eBay:iPads:: sold classification:: ebayipads_txt2_n15
bdanalytics
Date: (Tue) Nov 10, 2015
Introduction:
Data: Source: Training: https://inclass.kaggle.com/c/15-071x-the-analytics-edge-summer-2015/download/eBayiPadTrain.csv
New: https://inclass.kaggle.com/c/15-071x-the-analytics-edge-summer-2015/download/eBayiPadTest.csv
Time period:
Synopsis:
Based on analysis utilizing <> techniques,
Regression results: First run:
Classification results: template: prdline.my == “Unknown” -> 296 Low.cor.X.glm: Leaderboard: 0.83458 -> Rank 288 / 1884 0.85514 newobs_tbl=[N=471, Y=327]; submit_filename=template_Final_glm_submit.csv OOB_conf_mtrx=[YN=125, NY=76]=201; max.Accuracy.OOB=0.7710; opt.prob.threshold.OOB=0.6 startprice=100.00; biddable=95.42; productline=49.22; D.T.like=29.75; D.T.use=26.32; D.T.box=21.53;
prdline: -> Worse than template prdline.my == “Unknown” -> 285 All.X.no.rnorm.rf: Leaderboard: 0.82649 newobs_tbl=[N=485, Y=313]; submit_filename=prdline_Final_rf_submit.csv OOB_conf_mtrx=[YN=119, NY=80]=199; max.Accuracy.OOB=0.8339; opt.prob.threshold.OOB=0.5 startprice=100.00; biddable=84.25; D.sum.TfIdf=7.28; D.T.use=4.26; D.T.veri=2.78; D.T.scratch=1.99; D.T.box=; D.T.like=; Low.cor.X.glm: Leaderboard: 0.81234 newobs_tbl=[N=471, Y=327]; submit_filename=prdline_Low_cor_X_glm_submit.csv OOB_conf_mtrx=[YN=125, NY=74]=199; max.Accuracy.OOB=0.8205; opt.prob.threshold.OOB=0.6 startprice=100.00; biddable=96.07; prdline.my=51.37; D.T.like=29.39; D.T.use=25.43; D.T.box=22.27; D.T.veri=; D.T.scratch=;
oobssmpl: -> Low.cor.X.glm: Leaderboard: 0.83402 newobs_tbl=[N=440, Y=358]; submit_filename=oobsmpl_Final_glm_submit OOB_conf_mtrx=[YN=114, NY=84]=198; max.Accuracy.OOB=0.7780; opt.prob.threshold.OOB=0.5 startprice=100.00; biddable=93.87; prdline.my=60.48; D.sum.TfIdf=; D.T.condition=8.69; D.T.screen=7.96; D.T.use=7.50; D.T.veri=; D.T.scratch=;
category: -> Low.cor.X.glm: Leaderboard: 0.82381 newobs_tbl=[N=470, Y=328]; submit_filename=category_Final_glm_submit OOB_conf_mtrx=[YN=119, NY=57]=176; max.Accuracy.OOB=0.8011; opt.prob.threshold.OOB=0.6 startprice=100.00; biddable=79.19; prdline.my=55.22; D.sum.TfIdf=; D.T.ipad=27.05; D.T.like=21.44; D.T.box=20.67; D.T.condition=; D.T.screen=;
dataclns: -> All.X.no.rnorm.rf: Leaderboard: 0.82211 newobs_tbl=[N=485, Y=313]; submit_filename=dataclns_Final_rf_submit OOB_conf_mtrx=[YN=104, NY=75]=179; max.Accuracy.OOB=0.7977; opt.prob.threshold.OOB=0.5 startprice.log=100.00; biddable=65.85; prdline.my=7.74; D.sum.TfIdf=; D.T.use=2.01; D.T.condition=1.87; D.T.veri=1.62; D.T.ipad=; D.T.like=; Low.cor.X.glm: Leaderboard: 0.79264 newobs_tbl=[N=460, Y=338]; submit_filename=dataclns_Low_cor_X_glm_submit OOB_conf_mtrx=[YN=113, NY=74]=187; max.Accuracy.OOB=0.7977; opt.prob.threshold.OOB=0.5 -> different from prev run of 0.6 biddable=100.00; startprice.log=91.85; prdline.my=38.34; D.sum.TfIdf=; D.T.ipad=29.92; D.T.box=27.76; D.T.work=25.79; D.T.use=; D.T.condition=;
txtterms: -> top_n = c(10) Low.cor.X.glm: Leaderboard: 0.81448 newobs_tbl=[N=442, Y=356]; submit_filename=txtterms_Final_glm_submit OOB_conf_mtrx=[YN=113, NY=69]=182; max.Accuracy.OOB=0.7943; opt.prob.threshold.OOB=0.5 biddable=100.00; startprice.log=90.11; prdline.my=37.65; D.sum.TfIdf=; D.T.ipad=28.67; D.T.work=24.90; D.T.great=21.44; # [1] “D.T.condit” “D.T.condition” “D.T.good” “D.T.ipad” “D.T.new”
# [6] “D.T.scratch” “D.T.screen” “D.T.this” “D.T.use” “D.T.work”
All.X.glm: Leaderboard: 0.81016
newobs_tbl=[N=445, Y=353]; submit_filename=txtterms_Final_glm_submit
OOB_conf_mtrx=[YN=108, NY=72]=180; max.Accuracy.OOB=0.7966;
opt.prob.threshold.OOB=0.5
biddable=100.00; startprice.log=88.24; prdline.my=33.81; D.sum.TfIdf=;
D.T.scratch=25.51; D.T.use=18.97; D.T.good=16.37; [1] “D.T.condit” “D.T.use” “D.T.scratch” “D.T.new” “D.T.good” “D.T.screen” [7] “D.T.great” “D.T.excel” “D.T.work” “D.T.ipad”
Max.cor.Y.rpart: Leaderboard: 0.79258
newobs_tbl=[N=439, Y=359]; submit_filename=txtterms_Final_rpart_submit
OOB_conf_mtrx=[YN=105, NY=76]=181; max.Accuracy.OOB=0.7954802;
opt.prob.threshold.OOB=0.5
startprice.log=100; biddable=; prdline.my=; D.sum.TfIdf=;
D.T.scratch=; D.T.use=; D.T.good=; [1] “D.T.condit” “D.T.use” “D.T.scratch” “D.T.new” “D.T.good” “D.T.screen” [7] “D.T.ipad” “D.T.great” “D.T.work” “D.T.excel”
All.X.no.rnorm.rf: Leaderboard: 0.80929
newobs_tbl=[N=545, Y=253]; submit_filename=txtterms_Final_rf_submit
OOB_conf_mtrx=[YN=108, NY=61]=169; max.Accuracy.OOB=0.8090395
opt.prob.threshold.OOB=0.5
startprice.log=100.00; biddable=78.82; idseq.my=63.43; prdline.my=45.57;
D.T.use=2.76; D.T.condit=2.35; D.T.scratch=2.00; D.T.good=; [1] “D.T.condit” “D.T.use” “D.T.scratch” “D.T.new” “D.T.good” “D.T.screen” [7] “D.T.ipad” “D.T.great” “D.T.work” “D.T.excel”
txtclstr: All.X.no.rnorm.rf: Leaderboard: 0.79363 -> 0.79573 newobs_tbl=[N=537, Y=261]; submit_filename=txtclstr_Final_rf_submit OOB_conf_mtrx=[YN=104, NY=61]=165; max.Accuracy.OOB=0.8135593 opt.prob.threshold.OOB=0.5 startprice.log=100.00; biddable=79.99; idseq.my=64.94; prdline.my=4.14; prdline.my.clusterid=1.15; [1] “D.T.condit” “D.T.use” “D.T.scratch” “D.T.new” “D.T.good” “D.T.screen” [7] “D.T.ipad” “D.T.great” “D.T.work” “D.T.excel”
dupobs: All.X.no.rnorm.rf: Leaderboard: 0.79295 newobs_tbl=[N=541, Y=257]; submit_filename=dupobs_Final_rf_submit OOB_conf_mtrx=[YN=114, NY=65]=179; max.Accuracy.OOB=0.7977401 opt.prob.threshold.OOB=0.5 startprice.log=100.00; biddable=94.49; idseq.my=67.40; prdline.my=4.48; prdline.my.clusterid=1.99; [1] “D.T.condit” “D.T.use” “D.T.scratch” “D.T.new” “D.T.good” “D.T.screen” [7] “D.T.ipad” “D.T.great” “D.T.work” “D.T.excel”
All.X.no.rnorm.rf: Leaderboard: 0.79652
newobs_tbl=[N=523, Y=275]; submit_filename=dupobs_Final_rf_submit
OOB_conf_mtrx=[YN=114, NY=65]=179; max.Accuracy.OOB=0.7977401
opt.prob.threshold.OOB=0.5
startprice.log=100.00; biddable=94.24; idseq.my=67.92;
prdline.my=4.33; prdline.my.clusterid=2.17; [1] “D.T.condit” “D.T.use” “D.T.scratch” “D.T.new” “D.T.good” “D.T.screen” [7] “D.T.ipad” “D.T.great” “D.T.work” “D.T.excel”
csmmdl: All.X.no.rnorm.rf: Leaderboard: 0.79396 newobs_tbl=[N=525, Y=273]; submit_filename=csmmdl_Final_rf_submit OOB_conf_mtrx=[YN=111, NY=66]=177; max.Accuracy.OOB=0.8000000 opt.prob.threshold.OOB=0.5 startprice.log=100.00; biddable=90.30; idseq.my=67.06; prdline.my=4.40; cellular.fctr=3.57; prdline.my.clusterid=2.08;
All.Interact.X.no.rnorm.rf: Leaderboard: 0.77867 newobs_tbl=[N=564, Y=234]; submit_filename=csmmdl_Final_rf_submit OOB_conf_mtrx=[YN=120, NY=53]=173; max.Accuracy.OOB=0.8045198 opt.prob.threshold.OOB=0.5 biddable=100.00; startprice.log=93.99; idseq.my=57.30; prdline.my=9.09; cellular.fctr=3.30; prdline.my.clusterid=2.35;
All.Interact.X.no.rnorm.rf: Leaderboard: 0.77152 newobs_tbl=[N=539, Y=259]; submit_filename=csmmdl_Final_rf_submit OOB_conf_mtrx=[YN=, NY=]=; max.Accuracy.OOB=0.8011299 opt.prob.threshold.OOB=0.5 biddable=100.00; startprice.log=94.93; idseq.my=57.12; prdline.my=9.29; cellular.fctr=3.20; prdline.my.clusterid=2.50; [1] “D.T.condit” “D.T.use” “D.T.scratch” “D.T.new” “D.T.good” “D.T.screen” [7] “D.T.ipad” “D.T.great” “D.T.work” “D.T.excel”
All.X.glmnet:
fit_RMSE=???; OOB_RMSE=115.1247; new_RMSE=115.1247;
prdline.my.fctr=100.00; condition.fctrNew=88.53; D.npnct09.log=84.34
biddable=16.48; idseq.my=57.27;spdiff:
All.Interact.X.no.rnorm.rf: Leaderboard: 0.78218 newobs_tbl=[N=517, Y=281]; submit_filename=spdiff_Final_rf_submit OOB_conf_mtrx=[YN=121, NY=38]=159; max.Accuracy.OOB=0.8203390 opt.prob.threshold.OOB=0.6 biddable=100.00; startprice.diff=57.53; idseq.my=41.31; prdline.my=11.43; cellular.fctr=2.36; prdline.my.clusterid=1.82;
All.X.no.rnorm.rf:
fit_RMSE=92.19; OOB_RMSE=130.86; new_RMSE=130.86;
biddable=100.00; prdline.my.fctr=61.92; idseq.my=57.77;
condition.fctr=29.53; storage.fctr=11.22; color.fctr=6.69;
cellular.fctr=6.11
All.X.no.rnorm.rf: Leaderboard: 0.77443
newobs_tbl=[N=606, Y=192]; submit_filename=spdiff_Final_rf_submit
OOB_conf_mtrx=[YN=112, NY=28]=140; max.Accuracy.OOB=0.8418079
opt.prob.threshold.OOB=0.6
startprice.diff=100.00; biddable=96.53; idseq.my=38.10;
prdline.my=3.65; cellular.fctr=2.21; prdline.my.clusterid=0.91; [1] “D.T.condit” “D.T.use” “D.T.scratch” “D.T.new” “D.T.good” “D.T.screen” [7] “D.T.ipad” “D.T.great” “D.T.work” “D.T.excel”
color: All.Interact.X.glmnet: fit_RMSE=88.64520; prdline.my.fctr:D.TfIdf.sum.stem.stop.Ratio=100.00; prdline.my.fctr:condition.fctr=77.35 D.TfIdf.sum.stem.stop.Ratio=68.18 prdline.my.fctr:color.fctr=68.12 prdline.my.fctr:storage.fctr=63.32
All.X.no.rnorm.rf: Leaderboard: 0.80638
newobs_tbl=[N=550, Y=248]; submit_filename=color_Final_rf_submit
OOB_conf_mtrx=[YN=108, NY=54]=162; max.Accuracy.OOB=0.8169492
opt.prob.threshold.OOB=0.5
biddable=100.00; startprice.diff=77.90; idseq.my=48.49;
D.ratio.sum.TfIdf.nwrds=6.48; storage.fctr=4.74;
D.TfIdf.sum.stem.stop.Ratio=4.57; prdline.my=4.32;[1] “D.T.condit” “D.T.use” “D.T.scratch” “D.T.new” “D.T.good” “D.T.screen” [7] “D.T.ipad” “D.T.great” “D.T.work” “D.T.excel”
All.Interact.X.no.rnorm.rf: Leaderboard: 0.72974
newobs_tbl=[N=682, Y=116]; submit_filename=assctxt_Final_rf_submit
OOB_conf_mtrx=[YN=125, NY=43]=168; max.Accuracy.OOB=0.8101695; max.auc.OOB=???;
opt.prob.threshold.OOB=0.6
biddable=100.00; startprice.diff=51.04; idseq.my=29.51;
startprice.diff:biddable=28.70
prdline.my.fctriPadmini:idseq.my=6.89
Highest max.auc.OOB=???; for model:
All.X.no.rnorm.rf: min.RMSE.fit=1.4967021biddable 100.00000000 idseq.my 98.00292371 startprice.unit9 34.31130220 prdl.my.descr.fctr 18.10984741 D.ratio.sum.TfIdf.nwrds 15.23549621 color.fctrUnknown 14.05520993 D.TfIdf.sum.stem.stop.Ratio 13.00884673 D.ratio.nstopwrds.nwrds 10.51165302
All.X.gbm: Leaderboard: 0.75430
newobs_tbl=[N=582, Y=216]; submit_filename=mdlsel_Final_gbm_submit
OOB_conf_mtrx=[YN=58, NY=65]=123;
max.Accuracy.OOB=0.8617978; max.auc.OOB=0.9367161;
opt.prob.threshold.OOB=0.5startprice.diff 100.0000000 100.00000000 biddable 66.6475055 65.40764971 idseq.my 1.8632456 4.55963698
splogdiff: All.X.gbm: Leaderboard: 0.70111 newobs_tbl=[N=553, Y=245]; submit_filename=splogdiff_Final_gbm_submit OOB_conf_mtrx=[YN=35, NY=101]=136; max.Accuracy.OOB=0.8471910; max.auc.OOB=0.9388912; opt.prob.threshold.OOB=0.3 startprice.log.diff 100.0000000 100.0000000 biddable 86.8563123 88.0261866 idseq.my 8.3580281 2.9054298
Forum Ideas: I then focused on feature engineering, each new variable brought its own little improvement so in the end i just kept adding new ones and let the models do their thing. Here are some i used: model (productline:storage:condition), isNew, model2 (product:isNew), 50 common words from descr, descrLength, capsFactor (% of caps in description), number of cheaper items of same model2, number of dearer items of same model2, priceFactor (vs. mean of price for model), priceFactor2 (vs. mean of price for model2), bigID (if ID> 11000 because there seems to be a huge drop in sales after some time), timeline (year of product launch, reasoning is you want to spend less money on older products).
Get the median startprice for each level of productline and condition. Take the difference from startprice as a new variable. I find median works much better than the mean since startprice is not normally distributed. I also created another binary variable on whether this difference is positive or negative.
Square root startprice
scale and center all the variables except sold, including the dummies.
Prediction Accuracy Enhancement Options:
- import.data chunk:
- which obs should be in fit vs. OOB (currently dirty.0 vs .1 is split 50%)
- inspect.data chunk:
- For date variables
- Appropriate factors ?
- Different / More last* features ?
- For date variables
- scrub.data chunk:
- transform.data chunk:
- derive features from multiple features
- manage.missing.data chunk:
- Not fill missing vars
- Fill missing numerics with a different algorithm
- Fill missing chars with data based on clusters
- extract.features chunk:
- Text variables: move to date extraction chunk ???
- Mine acronyms
- Mine places
- Text variables: move to date extraction chunk ???
- Review set_global_options chunk after features are finalized
[](.png)
Potential next steps include:
- Organization:
- Categorize by chunk
- Priority criteria:
- Ease of change
- Impacts report
- Cleans innards
- Bug report
- all chunks:
- at chunk-end rm(!glb_)
- manage.missing.data chunk:
- cleaner way to manage re-splitting of training vs. new entity
- extract.features chunk:
- Add n-grams for glbFeatsText
- “RTextTools”, “tau”, “RWeka”, and “textcat” packages
- Convert user-specified mutate code to config specs
- Add n-grams for glbFeatsText
- fit.models chunk:
- Prediction accuracy scatter graph:
- Add tiles (raw vs. PCA)
- Use shiny for drop-down of “important” features
Use plot.ly for interactive plots ?
- Change .fit suffix of model metrics to .mdl if it’s data independent (e.g. AIC, Adj.R.Squared - is it truly data independent ?, etc.)
- move model_type parameter to myfit_mdl before indep_vars_vctr (keep all model_* together)
- create a custom model for rpart that has minbucket as a tuning parameter
varImp for randomForest crashes in caret version:6.0.41 -> submit bug report
- Probability handling for multinomials vs. desired binomial outcome
- ROCR currently supports only evaluation of binary classification tasks (version 1.0.7)
extensions toward multiclass classification are scheduled for the next release
- Skip trControl.method=“cv” for dummy classifier ?
- Add custom model to caret for a dummy (baseline) classifier (binomial & multinomial) that generates proba/outcomes which mimics the freq distribution of glb_rsp_var values; Right now glb_dmy_glm_mdl always generates most frequent outcome in training data
glm_dmy_mdl should use the same method as glm_sel_mdl until custom dummy classifer is implemented
- fit.all.training chunk:
- myplot_prediction_classification: displays ‘x’ instead of ‘+’ when there are no prediction errors
- Compare glb_sel_mdl vs. glb_fin_mdl:
- varImp
- Prediction differences (shd be minimal ?)
- Move glb_analytics_diag_plots to mydsutils.R: (+) Easier to debug (-) Too many glb vars used
- Add print(ggplot.petrinet(glb_analytics_pn) + coord_flip()) at the end of every major chunk
- Parameterize glb_analytics_pn
- Move glb_impute_missing_data to mydsutils.R: (-) Too many glb vars used; glb_<>_df reassigned
- Replicate myfit_mdl_classification features in myfit_mdl_regression
- Do non-glm methods handle interaction terms ?
- f-score computation for classifiers should be summation across outcomes (not just the desired one ?)
- Add accuracy computation to glb_dmy_mdl in predict.data.new chunk
- Why does splitting fit.data.training.all chunk into separate chunks add an overhead of ~30 secs ? It’s not rbind b/c other chunks have lower elapsed time. Is it the number of plots ?
- Incorporate code chunks in print_sessionInfo
- Test against
- projects in github.com/bdanalytics
- lectures in jhu-datascience track
Analysis:
rm(list = ls())
set.seed(12345)
options(stringsAsFactors = FALSE)
source("~/Dropbox/datascience/R/myscript.R")
source("~/Dropbox/datascience/R/mydsutils.R")## Loading required package: caret
## Loading required package: lattice
## Loading required package: ggplot2source("~/Dropbox/datascience/R/myplot.R")
source("~/Dropbox/datascience/R/mypetrinet.R")
source("~/Dropbox/datascience/R/myplclust.R")
source("~/Dropbox/datascience/R/mytm.R")
# Gather all package requirements here
suppressPackageStartupMessages(require(doMC))
registerDoMC(6) # # of cores on machine - 2
suppressPackageStartupMessages(require(caret))
source("~/Documents/Work/PullRequests/caret/pkg/caret/R/confusionMatrix.R")
source("~/Documents/Work/PullRequests/caret/pkg/caret/R/ggplot.R")
#packageVersion("tm")
#require(sos); findFn("cosine", maxPages=2, sortby="MaxScore")
# Analysis control global variables
# Inputs
glb_trnng_url <- "https://inclass.kaggle.com/c/15-071x-the-analytics-edge-summer-2015/download/eBayiPadTrain.csv"
glb_newdt_url <- "https://inclass.kaggle.com/c/15-071x-the-analytics-edge-summer-2015/download/eBayiPadTest.csv"
glbInpMerge <- # NULL #: default
list(fnames = c("ebayipads_finmdl_bid0_sp_out.csv", "ebayipads_mdlens_bid1_sp_out.csv"))
glb_is_separate_newobs_dataset <- TRUE # or TRUE
glb_split_entity_newobs_datasets <- FALSE # select from c(FALSE, TRUE)
glb_split_newdata_method <- NULL # select from c(NULL, "condition", "sample", "copy")
glb_split_newdata_condition <- NULL # or "is.na(<var>)"; "<var> <condition_operator> <value>"
glb_split_newdata_size_ratio <- 0.3 # > 0 & < 1
glb_split_sample.seed <- 123 # or any integer
glbObsDropCondition <- # default : NULL
"(UniqueID %in% c(NULL
, 11234 #sold=0; 2 other dups(10306, 11503) are sold=1
, 11844 #sold=0; 3 other dups(11721, 11738, 11812) are sold=1
)) |
(productline %in% c('iPad 5', 'iPad mini Retina'))
# | (biddable != 0) # bid0_sp
# | (biddable == 0) # bid1_sp
"
#parse(text=glbObsDropCondition)
#subset(glb_allobs_df, .grpid %in% c(31))
glb_obs_repartition_train_condition <- NULL
# "!is.na(sold) & (sold == 1)" # : bid._sp
glb_max_fitobs <- NULL # or any integer
glb_is_regression <- FALSE; glb_is_classification <- !glb_is_regression;
glb_is_binomial <- TRUE #or FALSE
glb_rsp_var_raw <- "sold" #: !_sp # "startprice" # : bid._sp #
# for classification, the response variable has to be a factor
glb_rsp_var <- "sold.fctr" #:!_sp # "startprice.log10" :bid._sp # glb_rsp_var_raw :default
# if the response factor is based on numbers/logicals e.g (0/1 OR TRUE/FALSE vs. "A"/"B"),
# or contains spaces (e.g. "Not in Labor Force")
# caret predict(..., type="prob") crashes
glb_map_rsp_raw_to_var <- function(raw) { # NULL
# return(raw ^ 0.5)
# return(log(1 + raw))
# return(log10(raw)) # bid._sp
# return(exp(-raw / 2))
ret_vals <- rep_len(NA, length(raw)); ret_vals[!is.na(raw)] <- ifelse(raw[!is.na(raw)] == 1, "Y", "N"); return(relevel(as.factor(ret_vals), ref="N"))
# #as.factor(paste0("B", raw))
# #as.factor(gsub(" ", "\\.", raw))
}
# glb_map_rsp_raw_to_var(tst <- c(NA, 0, 1)) # !_sp
# glb_map_rsp_raw_to_var(tst <- c(NA, 0, 2.99, 280.50, 1000.00)) # bid._sp
glb_map_rsp_var_to_raw <- function(var) { # NULL #
# return(var ^ 2.0)
# return(exp(var) - 1)
# return(10 ^ var) # bid._sp
# return(-log(var) * 2)
as.numeric(var) - 1
# #as.numeric(var)
# #gsub("\\.", " ", levels(var)[as.numeric(var)])
# c("<=50K", " >50K")[as.numeric(var)]
# #c(FALSE, TRUE)[as.numeric(var)]
}
# glb_map_rsp_var_to_raw(glb_map_rsp_raw_to_var(tst))
if ((glb_rsp_var != glb_rsp_var_raw) && is.null(glb_map_rsp_raw_to_var))
stop("glb_map_rsp_raw_to_var function expected")
glb_rsp_var_out <- paste0(glb_rsp_var, ".predict.") # mdl_id is appended later
# List info gathered for various columns
# <col_name>: <description>; <notes>
# description = The text description of the product provided by the seller.
# biddable = Whether this is an auction (biddable=1) or a sale with a fixed price (biddable=0).
# startprice = The start price (in US Dollars) for the auction (if biddable=1) or the sale price (if biddable=0).
# condition = The condition of the product (new, used, etc.)
# cellular = Whether the iPad has cellular connectivity (cellular=1) or not (cellular=0).
# carrier = The cellular carrier for which the iPad is equipped (if cellular=1); listed as "None" if cellular=0.
# color = The color of the iPad.
# storage = The iPad's storage capacity (in gigabytes).
# productline = The name of the product being sold.
# If multiple vars are parts of id, consider concatenating them to create one id var
# If glb_id_var == NULL, ".rownames <- row.names()" is the default
# Derive a numeric feature from id var
# User-specified exclusions
# List feats that shd be excluded due to known causation by prediction variable
glbFeatsExclude <- c(NULL
### !_sp
, "description", "descr.my", "productline"
, "startprice", "startprice.log10.predict", "sprice.predict.diff"
### bid0_sp
# , "description", "productline"
# , "sold", "startprice.log10.cut.fctr"
# # List feats that are linear combinations (alias in glm)
# , "D.terms.post.stem.n.log", "D.weight.sum"
# #, "prdl.descr.my.fctriPad4#1:.clusterid.fctr3" This does not work
# # if RFE is rated lower than Low.cor, list feats that are in RFE & not in Low.cor
# # min.RMSE.fit(RFE.X.glmnet)=0.1138888
# # D.chrs.n.log 61.12483
# # D.chrs.uppr.n.log 61.12483
# # D.ratio.wrds.stop.n.wrds.n 61.12483
# # D.terms.post.stop.n.log 61.12483
# # D.weight.post.stem.sum 61.12483
# # D.wrds.n.log 61.12483
# # D.wrds.stop.n.log 61.12483
# # D.wrds.unq.n.log 61.12483
# #, "startprice.dcm2.is9" # min.RMSE.fit(RFE.X.glmnet)=0.1141991 (up)
# , "D.wrds.stop.n.log" # min.RMSE.fit(RFE.X.glmnet)=0.1131232
### bid0_sp
### bid1_sp
# , "description", "productline"
# , "sold", "startprice.log10.cut.fctr"
### bid1_sp
)
glb_id_var <- c("UniqueID")
glb_category_var <- "prdl.my.fctr" # "prdl.descr.my.fctr" # "productline" # NULL
glb_drop_vars <- c(NULL) # or c("<col_name>")
glb_map_vars <- NULL # or c("<var1>", "<var2>")
glb_map_urls <- list();
# glb_map_urls[["<var1>"]] <- "<var1.url>"
glb_assign_pairs_lst <- NULL;
# glb_assign_pairs_lst[["<var1>"]] <- list(from=c(NA),
# to=c("NA.my"))
glb_assign_vars <- names(glb_assign_pairs_lst)
# Derived features
glbFeatsDerive <- NULL;
# Add logs of numerics that are not distributed normally -> do automatically ???
# Right skew: logp1; sqrt; ^ 1/3; logp1(logp1); log10; exp(-<feat>/constant)
# glbFeatsDerive[["prdline.my"]] <- list(
# mapfn=function(productline) { return(productline) }
# , args=c("productline"))
### bid._sp
# glbFeatsDerive[["startprice.log10.cut.fctr"]] <- list(
# mapfn=function(startprice.log10) { return(cut(startprice.log10, 3)) }
# , args=c("startprice.log10"))
### bid._sp
glbFeatsDerive[["sprice.root2"]] <- list(
mapfn = function(startprice) { return(startprice ^ (1/2)) }
, args = c("startprice"))
glbFeatsDerive[["sprice.log10"]] <- list(
mapfn = function(startprice) { return(log(startprice)) }
, args = c("startprice"))
glbFeatsDerive[["sprice.d20nexp"]] <- list(
mapfn = function(startprice) { return(exp(-startprice / 20)) }
, args = c("startprice"))
glbFeatsDerive[["sprice.predict.diff"]] <- list(
mapfn = function(startprice.log10.predict, startprice) {
spdiff <- (10 ^ startprice.log10.predict) - startprice;
return(spdiff) }
, args = c("startprice.log10.predict", "startprice"))
# glbFeatsDerive[["spdiff.root10"]] <- list(
# mapfn = function(sprice.predict.diff) {
# return(sign(sprice.predict.diff) * (abs(sprice.predict.diff) ^ (1/10))) }
# , args = c("sprice.predict.diff"))
glbFeatsDerive[["spdiff.cut.fctr"]] <- list(
mapfn = function(sprice.predict.diff) {
return(cut(sprice.predict.diff, c(-1000, -100, -10, -1, 0, 1, 10, 100, 1000))) }
, args = c("sprice.predict.diff"))
#glb_allobs_df[which(glb_post_stop_words_terms_mtrx_lst[[txt_var]][, subset(glb_post_stop_words_terms_df_lst[[txt_var]], term %in% c("conditionminimal"))$pos] > 0), "description"]
glbFeatsDerive[["descr.my"]] <- list(
mapfn = function(description) { mod_raw <- description;
### bid._sp
# # This is here because it does not work with txt_map_filename
mod_raw <- gsub(paste0(c("\n", "\211", "\235", "\317", "\333"), collapse = "|"), " ",
mod_raw)
# # This should go into txt_map_filename
# mod_raw <- gsub("\\.\\.", "\\. ", mod_raw);
# # Don't parse for "." because of ".com"; use customized gsub for that text
# mod_raw <- gsub("(\\w)(!|\\*|,|-|/)(\\w)", "\\1\\2 \\3", mod_raw);
#mod_raw <- grep(""", glb_allobs_df$descr.my, value = TRUE)
mod_raw <- gsub("&", "&", mod_raw);
mod_raw <- gsub("<", "<", mod_raw);
mod_raw <- gsub(">", ">", mod_raw);
mod_raw <- gsub("<br>", " ", mod_raw); # line break - add a count for it ???
mod_raw <- gsub(""", " ", mod_raw); # justification meta-character
mod_raw <- gsub("&#(0*)37;", "%", mod_raw);
mod_raw <- gsub("'", "'", mod_raw);
mod_raw <- gsub("([[:digit:]])\\.([[:digit:]])\\.([[:digit:]])",
"\\1point\\2\\point\\3", mod_raw);
mod_raw <- gsub("([[:digit:]])\\.([[:digit:]])", "\\1point\\2", mod_raw);
mod_raw <- gsub("([[:digit:]]),([[:digit:]])", "\\1\\2", mod_raw);
mod_raw <- gsub("\\b1st\\b", "first", mod_raw);
mod_raw <- gsub("\\b2nd\\b", "second", mod_raw);
mod_raw <- gsub("\\b3rd\\b", "third", mod_raw);
mod_raw <- gsub("\\b4th\\b", "fourth", mod_raw);
mod_raw <- gsub("\\.(com|COM)\\b", "dot\\1", mod_raw);
#
# # Modifications for this exercise only
# # Add dictionary to stemDocument e.g. stickers stemmed to sticker ???
# mod_raw <- gsub("8\\.25", "825", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\b10\\.SCREEN\\b", "10\\. SCREEN", mod_raw);
mod_raw <- gsub("\\b128 gb\\b", "128gb", mod_raw);
mod_raw <- gsub("\\b16G\\b", "16GB", mod_raw);
# mod_raw <- gsub(" 16 gig ", " 16gb ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" 16 gb ", " 16gb ", mod_raw, ignore.case=TRUE);
#
# mod_raw <- gsub("\\bAccounts\\b", "Account", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\bactivated\\b", "activate", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub(" actuuly ", " actual ", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\badaptor\\b", "adapter", mod_raw);
# mod_raw <- gsub("\\baffects\\b", "affect", mod_raw, ignore.case=FALSE);
mod_raw <- gsub("\\bair-like\\b", "air -like", mod_raw);
mod_raw <- gsub("\\bALL-JUST\\b", "ALL -JUST", mod_raw);
mod_raw <- gsub("\\bApple's\\b", "Apple'", mod_raw);
# #mod_raw <- glb_allobs_df[c(1322), txt_var]; mod_raw
mod_raw <- gsub("\\bApple care\\b", "Applecare", mod_raw);
mod_raw <- gsub("\\bAT&T\\b", "ATT", mod_raw);
# mod_raw <- gsub(" bacK!wiped ", " bacK ! wiped ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" backplate", " back plate", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bbarley", "barely", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" bend ", " bent ", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\b(B|b)(EST|est) (B|b)(UY|uy)\\b", "\\1\\2\\3\\4", mod_raw);
# mod_raw <- gsub(" black\\.Device ", " black \\. Device ", mod_raw,
# ignore.case=TRUE);
# mod_raw <- gsub("black\\),charger ", "black\\), charger ", mod_raw,
# ignore.case=TRUE);
# mod_raw <- gsub("\\bblacked\\b", "black", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\bblemish\\b", "blemishes", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub(" blocks", " blocked", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" book ", " manual ", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\b(B|b)(RAND|rand)( |-)(N|n)(EW|ew)\\b", "\\1\\2\\4\\5", mod_raw)
#mod_raw <- c("brand new", "BRAND new", "brand NEW", "BRAND NEW", "bbrand new", "brand-new", "brand newb")
mod_raw <- gsub("\\bbrokenCharger\\b", "broken Charger", mod_raw);
#
mod_raw <- gsub("\\bC-Major\\b", "C -Major", mod_raw)
# mod_raw <- gsub(" perfectlycord ", " perfectly cord ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bcord", "cable", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\bcables\\.No\\b", "cables. No", mod_raw);
# mod_raw <- gsub("\\bcables\\b", "cable", mod_raw, ignore.case=TRUE);
#
mod_raw <- gsub("\\bcare\\.The\\b", "care\\. The", mod_raw);
# mod_raw <- gsub("\\b(cared|careful|CAREFUL)\\b", "care", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\b(cases|casing)\\b", "case", mod_raw, ignore.case=TRUE);
# #mod_raw <- glb_allobs_df[c(88,187,280,1040,1098), txt_var]; mod_raw
mod_raw <- gsub("\\bCase/Cover\\b", "Case/ Cover", mod_raw);
mod_raw <- gsub("\\bCasing/Screen\\b", "Casing/ Screen", mod_raw);
# mod_raw <- gsub(" carefully ", " careful ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bchargers\\b", "charger", mod_raw, ignore.case=FALSE);
mod_raw <- gsub("\\bchip/crack\\b", "chip/ crack", mod_raw);
# mod_raw <- gsub("\\bchips\\b", "chip", mod_raw, ignore.case=FALSE);
mod_raw <- gsub("\\bCLEAN\\!LIKE\\b", "CLEAN! LIKE", mod_raw);
# mod_raw <- gsub("\\bcleanly\\b", "clean", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\b(C|c)olor(.*)s\\b", "\\1olor", mod_raw, ignore.case=FALSE);
# #mod_raw <- glb_allobs_df[c(280,1411), txt_var]; mod_raw
mod_raw <- gsub("\\bColors,models\\b", "Colors ,models", mod_raw);
# mod_raw <- gsub("\\bcompletely\\b", "complete", mod_raw, ignore.case=FALSE);
# #mod_raw <- glb_allobs_df[c(178), txt_var]; mod_raw
#
mod_raw <- gsub("\\bCONDITION..CLEAN\\b", "CONDITION ..CLEAN", mod_raw);
mod_raw <- gsub("\\bcondition,comes\\b", "condition ,comes", mod_raw);
mod_raw <- gsub("\\bcondition\\.Device\\b", "condition .Device", mod_raw);
mod_raw <- gsub("\\bconditionHas\\b", "condition Has", mod_raw);
mod_raw <- gsub("\\bcondition\\.\\.\\.like\\b", "condition ...like", mod_raw);
mod_raw <- gsub("\\bcondition\\*Minimal\\b", "condition *Minimal", mod_raw);
mod_raw <- gsub("\\bCondition-Moderate\\b", "Condition -Moderate", mod_raw);
mod_raw <- gsub("\\bcondition\\.The\\b", "condition .The", mod_raw);
mod_raw <- gsub("\\bCONDITION\\.VERY\\b", "CONDITION .VERY", mod_raw);
# mod_raw <- gsub(" (conditon|condtion|contidion|conditions)", " condition", mod_raw,
# mod_raw <- gsub("\\b(conditon|condtion|contidion|conditions)\\b", "condition", mod_raw,
# ", "\\1\\. \\2", mod_raw,
# ignore.case=TRUE);
# mod_raw <- gsub("(condition)(Has)", "\\1\\. \\2", mod_raw);
#
# mod_raw <- gsub("\\bCONNECTED\\b", "CONNECT", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\bconnects\\b", "connect", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub(" consist ", " consistent ", mod_raw, ignore.case=TRUE);
# #mod_raw <- glb_allobs_df[c(195, 379, 437), txt_var]; mod_raw
# mod_raw <- gsub("\\bCosmetics\\b", "Cosmetic", mod_raw, ignore.case=FALSE);
mod_raw <- gsub("\\bCracked/Damaged\\b", "Cracked/ Damaged", mod_raw);
mod_raw <- gsub("\\bcracksNo\\b", "cracks No", mod_raw);
#
# mod_raw <- gsub("\\b(D|d)amaged\\b", "\\1amage", mod_raw, ignore.case=TRUE);
# #mod_raw <- glb_allobs_df[c(116, 1360), txt_var]; mod_raw
# mod_raw <- gsub("\\bDays\\b", "Day", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" DEFAULTING ", " DEFAULT ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bdefect(ive)*\\b", "defects", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub(" definitely ", " definite ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\b(D|d)ented\\b", "\\1ent", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub(" described", " describe", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" desciption", " description", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" devices", " device", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" Digi\\.", " Digitizer\\.", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\b(ding|dinged)\\b", "dings", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" display\\.New ", " display\\. New ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" displays", " display", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\bdo( +)not\\b", "dont", mod_raw);
mod_raw <- gsub("\\b(D|d)oes( +)(N|n)(O|o)(T|t)\\b", "\\1oes\\3\\5", mod_raw);
# mod_raw <- gsub("\\b(drop|drops)\\b", "dropped", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\b(E|e)dge\\b", "\\1dges", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub(" effect ", " affect ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" Excellant ", " Excellent ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" excellently", " excellent", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" EUC ", " excellent used condition", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" feels ", " feel ", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\bfineiCloud\\b", "fine iCloud", mod_raw, ignore.case = FALSE);
# mod_raw <- gsub(" fine.Its ", " fine. Its ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bfix\\b", "fixed", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bflaws\\b", "flaw", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bflawlessly\\b", "flawless", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" Framing ", " Frame ", mod_raw, ignore.case=TRUE);
#
# mod_raw <- gsub(" functioanlity", " functionality", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bfunction(ing|ality)\\b", "functional", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" functional\\.Very little ", " functional\\. Very little ", mod_raw,
# ignore.case=TRUE);
mod_raw <- gsub("\\b([[:digit:]]+) (GB|gb)\\b", "\\1\\2", mod_raw);
mod_raw <- gsub("\\b([[:digit:]]+) gig\\b", "\\1gb", mod_raw);
mod_raw <- gsub("\\b(G|g)(EEK|eek) (S|s)(QUAD|quad)\\b", "\\1\\2\\3\\4", mod_raw);
# mod_raw <- gsub("^Gentle ", "Gently ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\(gray color", "\\(spacegray color", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" GREAT\\.SCreen ", " GREAT\\. SCreen ", mod_raw,
# ignore.case=TRUE);
mod_raw <- gsub("\\bGUARANTEES-IT\\b", "GUARANTEES -IT", mod_raw);
# mod_raw <- gsub("\\b(guarantee|guarantees)\\b", "guaranteed", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\ba handful of times\\b", "sparingly", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bhardly any\\b", "no", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bhardly ever used\\b", "sparingly used", mod_raw, ignore.case=TRUE);
#
mod_raw <- gsub("\\biCL0UD\\b", "iCLOUD", mod_raw);
mod_raw <- gsub("\\bI (CLOUD|cloud)\\b", "I\\1", mod_raw);
# mod_raw <- gsub("^iPad Black 3rd generation ", "iPad 3 Black ", mod_raw,
# ignore.case=TRUE);
mod_raw <- gsub("\\bIMEINo\\b", "IMEI No", mod_raw);
mod_raw <- gsub("\\bIMIE\\b", "IMEI", mod_raw);
# mod_raw <- gsub("\\bincluding\\b", "included", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub(" install\\. ", " installed\\. ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("inivisible", "invisible", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\bI pad\\b", "Ipad", mod_raw);
mod_raw <- gsub("\\b(I|i)(P|p)(A|a)(D|d) (A|a)(I|i)(R|r)\\b", "\\1\\2\\3\\4\\5\\6\\7",
mod_raw);
mod_raw <- gsub("\\b(I|i)(P|p)(A|a)(D|d) (M|m)ini\\b", "\\1\\2\\3\\4\\5ini", mod_raw);
mod_raw <- gsub("\\b(I|i)(P|p)(A|a)(D|d) (M|m)inis\\b", "\\1\\2\\3\\4\\5ini", mod_raw);
mod_raw <- gsub("\\b(IPAD|Ipad|iPad|ipad) ([[:digit:]])\\b", "\\1\\2", mod_raw);
mod_raw <- gsub("\\b(Ipadair|iPadAir|ipadair) ([[:digit:]])\\b", "\\1\\2",
mod_raw);
mod_raw <- gsub("\\b(iPadMini|iPadmini) ([[:digit:]])\\b", "\\1\\2", mod_raw);
mod_raw <- gsub("\\bI Phone\\b", "IPhone", mod_raw);
mod_raw <- gsub("\\bIS-NO\\b", "IS -NO", mod_raw, ignore.case = FALSE)
#
# mod_raw <- gsub(" Keeped ", " Kept ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" knicks ", " nicks ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" lightening ", " lightning ", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\bLightning-to-USB\\b", "Lightning- to- USB", mod_raw);
mod_raw <- gsub("\\b(L|l)(IKE|ike)( |-)(N|n)(EW|ew)\\b", "\\1\\2\\4\\5", mod_raw);
#mod_raw <- c("like new", "LIKE new", "like NEW", "LIKE NEW", "blike new", "like-new", "like newb")
mod_raw <- gsub("\\bLIKENEW!ONE\\b", "LIKENEW! ONE", mod_raw);
# mod_raw <- gsub("\\b(lock|locks)\\b", "locked", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\blots\\b", "lot", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" manuals ", " manual ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" mars ", " marks ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" marks\\.Absolutely ", " marks\\. Absolutely ", mod_raw,
# ignore.case=TRUE);
# mod_raw <- gsub("\\bmarkings\\b", "marks", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bminis\\b", "mini", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub(" minimum", " minimal", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" MINT\\.wiped ", " MINT\\. wiped ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bmonth\\b", "months", mod_raw, ignore.case=TRUE);
# #mod_raw <- glb_allobs_df[c(1803), txt_var]; mod_raw
mod_raw <- gsub("\\bNew-Other\\b", "New -Other", mod_raw);
# mod_raw <- gsub(" NEW\\!(SCREEN|ONE) ", " NEW\\! \\1 ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" new looking$", " looks new", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" newer ", " new ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bnoted\\b", "note", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" oped ", " opened ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" opening", " opened", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" operated", " operational", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\botter box\\b", "otterbox", mod_raw);
#
# mod_raw <- gsub("\\bpackage\\b", "packaging", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\bPACKAGE\\b", "PACKAGing", mod_raw, ignore.case=FALSE);
# #mod_raw <- glb_allobs_df[c(360, 1142), txt_var]; mod_raw
mod_raw <- gsub("\\bperfectlycord\\b", "perfectly cord", mod_raw);
# mod_raw <- gsub(" performance", " performs", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" personalized ", " personal ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bPhysically\\b", "Physical", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\b(picture|pictured)\\b", "pictures", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\bPICTURE\\b", "PICTUREs", mod_raw, ignore.case=FALSE);
# #mod_raw <- glb_allobs_df[c(184, 892), txt_var]; mod_raw
# mod_raw <- gsub("\\b[P|p]ower(ed|ing|s)\\b", "\\1ower", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub(" pre- owned ", " used ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bprevious\\b", "previously", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bpreviously (owned|used)\\b", "used", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\b(P|p)reviously(.*)(O|o)wned\\b", "\\1reviously\\3wned\\2",
mod_raw);
mod_raw <- gsub("\\b(P|p)reviously(.*)(U|u)sed\\b", "\\1reviously\\3sed\\2", mod_raw);
# mod_raw <- gsub("\\bproblem\\b", "problems", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" products ", " product ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bprotected\\b", "protector", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\bprotection\\b", "protector", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\bPROTECTION\\b", "PROTECTOR", mod_raw, ignore.case=FALSE);
mod_raw <- gsub("\\bREADiPad\\b", "READ iPad", mod_raw);
# mod_raw <- gsub(" re- assemble ", " reassemble ", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\bREFURB\\.", "REFURBished.", mod_raw);
# mod_raw <- gsub(" reponding", " respond", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\bright-hand\\b", "right -hand", mod_raw);
# mod_raw <- gsub(" rotation ", " rotate ", mod_raw, ignore.case=TRUE);
#
# mod_raw <- gsub(" Sales ", " Sale ", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\bScratch-Free\\b", "Scratch- Free", mod_raw);
mod_raw <- gsub("\\bSCRATCHES/BLEMISHES...SCRATCHES\\b", "SCRATCHES/ BLEMISHES... SCRATCHES", mod_raw);
mod_raw <- gsub("\\bscratches,clear\\b", "scratches, clear", mod_raw);
mod_raw <- gsub("\\bScratches/Dent\\b", "Scratches/ Dent", mod_raw);
mod_raw <- gsub("\\bScratches/scuffs/nicks/scrapes\\b", "Scratches/ scuffs/ nicks/ scrapes", mod_raw);
# mod_raw <- gsub(" scratchs ", " scratches ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\b(scratchs|scratching)\\b", "scratches", mod_raw, ignore.case=FALSE);
mod_raw <- gsub("\\bset up\\b", "setup", mod_raw);
# mod_raw <- gsub(" shipped| Shipment", " ship", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bshowing\\b", "shows", mod_raw, ignore.case=FALSE);
mod_raw <- gsub("\\b(shrink|SHRINK) (wrap|WRAP)", "\\1\\2", mod_raw);
# mod_raw <- gsub("\\bshuts\\b", "shut", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" sides ", " side ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" skinned,", " skin,", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bSlightly\\b", "slight", mod_raw, ignore.case=FALSE);
mod_raw <- gsub("\\b(Space|space) (G|g)r(a|e)y\\b", "\\1\\2ray", mod_raw);
# mod_raw <- gsub(" spec ", " speck ", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\bsomescratches\\b", "some scratches", mod_raw);
# mod_raw <- gsub(" Sticker ", " Stickers ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bstoring", "store", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("SWAPPA\\.COM", "SWAPPAsdotCOM", mod_raw, ignore.case=TRUE);
#
# mod_raw <- gsub(" T- Mobile", " TMobile", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\b(tear|TEAR)(s|S)\\b", "\\1", mod_raw, ignore.case=FALSE);
# #mod_raw <- glb_allobs_df[c(376), txt_var]; mod_raw
mod_raw <- gsub("\\b(touch|Touch|TOUCH) (screen|SCREEN)\\b", "\\1\\2", mod_raw);
# mod_raw <- gsub("\\bTURN\\b", "TURNS", mod_raw, ignore.case=FALSE);
#
mod_raw <- gsub("\\bUnlockedCracked\\b", "Unlocked Cracked", mod_raw);
# mod_raw <- gsub("\\bUNUSABLE\\b", "UNUSED", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\b(update|updates)\\b", "updated", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub("\\bupgrade\\b", "upgraded", mod_raw, ignore.case=FALSE);
# mod_raw <- gsub(" uppser ", " upper ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("use*Case\\b", "use *Case", mod_raw, ignore.case = FALSE)
# mod_raw <- gsub(" use\\.Scratches ", " use\\. Scratches ", mod_raw,
# ignore.case=TRUE);
#
# mod_raw <- gsub(" verify ", " verified ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" wear\\.Device ", " wear\\. Device ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub("\\bwears\\b", "\\wear", mod_raw, ignore.case=TRUE);
# #mod_raw <- glb_allobs_df[c(167, 272), txt_var]; mod_raw
# mod_raw <- gsub(" whats ", " what's ", mod_raw, ignore.case=TRUE);
# mod_raw <- gsub(" WiFi\\+4G ", " WiFi \\+ 4G ", mod_raw, ignore.case=TRUE);
mod_raw <- gsub("\\b(W|w)(IFI|ifi)( |-)(ONLY|only)\\b", "\\1\\2\\4", mod_raw);
mod_raw <- gsub("\\bwill( +)not\\b", "wont", mod_raw);
mod_raw <- gsub("\\byr\\b", "year", mod_raw);
mod_raw <- gsub("\\bZa(a|g)g Invisible(.*)Shield\\b", "ZaagInvisibleShield", mod_raw);
### bid._sp
return(mod_raw) }
, args = c("description"))
# To identify glb_id_vars with >=10 obs
#mod_raw <- glb_allobs_df[sel_obs(list(descr.my.contains="\\bdoes( +)not\\b")), glbFeatsText]
#mod_raw <- glb_allobs_df[sel_obs(list(descr.my.contains="\\bipad [[:digit:]]\\b")), glbFeatsText][01:10]
#mod_raw <- glb_allobs_df[sel_obs(list(descr.my.contains="pad mini")), glbFeatsText][11:20]
#mod_raw <- glb_allobs_df[sel_obs(list(descr.my.contains="pad mini")), glbFeatsText][21:30]
#mod_raw <- glb_allobs_df[sel_obs(list(descr.my.contains="pad mini")), glbFeatsText][31:40]
glbFeatsDerive[["prdl.my.fctr"]] <- list(
mapfn = function(productline, description) {
as.factor(gsub(" ", "", productline)) }
, args = c("productline"))
# glbFeatsDerive[["prdl.descr.my.fctr"]] <- list(
# mapfn = function(productline, description) {
# as.factor(paste(gsub(" ", "", productline),
# as.numeric(nchar(description) > 0),
# sep = "#")) }
# , args = c("productline", "description"))
#print(mycreate_sqlxtab_df(glb_allobs_df, c("prdl.descr.my.fctr", "sold")))
# mapfn=function(startprice) { return(scale(log(startprice))) }
# , args=c("startprice"))
# mapfn=function(Rasmussen) { return(ifelse(sign(Rasmussen) >= 0, 1, 0)) }
# mapfn=function(PropR) { return(as.factor(ifelse(PropR >= 0.5, "Y", "N"))) }
# mapfn=function(purpose) { return(relevel(as.factor(purpose), ref="all_other")) }
# mapfn=function(Week) { return(substr(Week, 1, 10)) }
# mapfn=function(raw) { tfr_raw <- as.character(cut(raw, 5));
# tfr_raw[is.na(tfr_raw)] <- "NA.my";
# return(as.factor(tfr_raw)) }
# , args=c("raw"))
# mapfn=function(PTS, oppPTS) { return(PTS - oppPTS) }
# , args=c("PTS", "oppPTS"))
# # If glb_allobs_df is not sorted in the desired manner
# mapfn=function(Week) { return(coredata(lag(zoo(orderBy(~Week, glb_allobs_df)$ILI), -2, na.pad=TRUE))) }
# mapfn=function(ILI) { return(coredata(lag(zoo(ILI), -2, na.pad=TRUE))) }
# mapfn=function(ILI.2.lag) { return(log(ILI.2.lag)) }
# glbFeatsDerive[["<txt_var>.niso8859.log"]] <- list(
# mapfn=function(<txt_var>) { match_lst <- gregexpr("&#[[:digit:]]{3};", <txt_var>)
# match_num_vctr <- unlist(lapply(match_lst,
# function(elem) length(elem)))
# return(log(1 + match_num_vctr)) }
# , args=c("<txt_var>"))
# mapfn=function(raw) { mod_raw <- raw;
# mod_raw <- gsub("&#[[:digit:]]{3};", " ", mod_raw);
# # Modifications for this exercise only
# mod_raw <- gsub("\\bgoodIn ", "good In", mod_raw);
# return(mod_raw)
# # Create user-specified pattern vectors
# #sum(mycount_pattern_occ("Metropolitan Diary:", glb_allobs_df$Abstract) > 0)
# if (txt_var %in% c("Snippet", "Abstract")) {
# txt_X_df[, paste0(txt_var_pfx, ".P.metropolitan.diary.colon")] <-
# as.integer(0 + mycount_pattern_occ("Metropolitan Diary:",
# glb_allobs_df[, txt_var]))
#summary(glb_allobs_df[ ,grep("P.on.this.day", names(glb_allobs_df), value=TRUE)])
# glb_allobs_df$<descriptor>.my <-
# plyr::revalue(glb_allobs_df$<descriptor>.my, c(
# "ABANDONED BUILDING" = "OTHER",
# "##" = "##"
# ))
# print(<descriptor>_freq_df <- mycreate_sqlxtab_df(glb_allobs_df, c("<descriptor>.my")))
# # print(dplyr::filter(<descriptor>_freq_df, grepl("(MEDICAL|DENTAL|OFFICE)", <descriptor>.my)))
# # print(dplyr::filter(dplyr::select(glb_allobs_df, -<var.zoo>),
# # grepl("STORE", <descriptor>.my)))
# glbFeatsDerive[["<var1>"]] <- glbFeatsDerive[["<var2>"]]
glb_derive_vars <- names(glbFeatsDerive)
# tst <- "descr.my"; args_lst <- NULL; for (arg in glbFeatsDerive[[tst]]$args) args_lst[[arg]] <- glb_allobs_df[, arg]; print(head(args_lst[[arg]])); print(head(drv_vals <- do.call(glbFeatsDerive[[tst]]$mapfn, args_lst)));
# print(which_ix <- which(args_lst[[arg]] == 0.75)); print(drv_vals[which_ix]);
glb_date_vars <- NULL # or c("<date_var>")
glb_date_fmts <- list(); #glb_date_fmts[["<date_var>"]] <- "%m/%e/%y"
glb_date_tzs <- list(); #glb_date_tzs[["<date_var>"]] <- "America/New_York"
#grep("America/New", OlsonNames(), value=TRUE)
glbFeatsPrice <- c("startprice") #: bid._sp # NULL or c("<price_var>")
# Text Processing Step: custom modifications not present in txt_munge
glbFeatsText <- c("descr.my") # NULL #
Sys.setlocale("LC_ALL", "C") # For english## [1] "C/C/C/C/C/en_US.UTF-8"# Text Processing Step: universal modifications
glb_txt_munge_filenames_pfx <- "ebay_mytxt_"
# Text Processing Step: tolower
# Text Processing Step: removePunctuation (use custom transformer to replace with space ???)
# Text Processing Step: removeWords
glb_txt_stop_words <- list()
# Remember to use unstemmed words; Check stemming of "significant" words - any stopped words that should be stemmed with them ?
if (!is.null(glbFeatsText)) {
require(tm)
glb_txt_stop_words[["descr.my"]] <- sort(c(NULL
### bid._sp
# , setdiff(removePunctuation(stopwords("english")), "no")
# ,"ac"
# # cor.y.train == NA
# ,unlist(strsplit(paste(c(NULL
# ,"128gb,1st,32gb,3g,64gb,90,acceptable,activation,amount,average,bad,buttons,buy,came,camera,can,care,carrier"
# #,casing
# ,"certified,charge,charging,cleaned,clear,come,components,contain,corner,correctly,covered,customer,earbuds"
# ,"engraved,engraving,engravement" # somehow didn't show up in the cor.y.train == NA list
# ,"entire,except,fair,features,feel,fine,generation,get,gift,got,heavily,heavy,however,imei,include,inspected,invisible,invisibleshield"
# ,"ipad,ipads"
# ,"issues"
# #,items,
# ,"keyboard,lightning,listing,little,looks,lower"
# ,"manufacture,manufacturer"# somehow didn't show up in the cor.y.train == NA list
# ,"meaning,model,near,need,needs,nicks,opened,operational,otherwise"
# ,"person,personal"# somehow didn't show up in the cor.y.train == NA list
# ,"phone,photos,pics,plastic,port,professionally"
# ,"purchased,purchasing"# somehow didn't show up in the cor.y.train == NA list
# ,"quality,questions,read,ready"
# ,"receive,received"# somehow didn't show up in the cor.y.train == NA list
# ,"removed,replaced,retail,return,returns,runs"
# #,scratch,
# ,"scuffing,sealed,sell,seller,selling,shape,ship,shown,silver,since,sold,sound,spacegray,stock,sync,tablet,taken,technician,tests,third,time,touch,units,unlocked,week,wifi,without"
# ,"wrap" # somehow didn't show up in the cor.y.train == NA list
# ,"zagg"
# ), collapse=",")
# , "[,]")) #err.abs.fit.sum=26.869473 w/o items,scratch
#
# # cor.y.abs is low
# #,"always","comes","grade","moderate","protector"
### bid._sp
### !_sp
# freq == 1; keep "gold"
,"09","17","28","34","360","5c","5point1point1","511","6428"
,"7point9","7point9in","79in"
,"8point5","8point25","82510"
,"910","9510","9point7","9point75","97510","99"
,"a1314","a1430","abused","accept","accounts","across","actuuly","add","advised"
,"affects","am","ans","antenna","anti","anyone","anything","applied","applying"
,"area","arizona","att","attached"
,"backlight","backlit","backplate","barley"
,"beetle","beginning","bend","besides","between"
,"bidder","binder","bonus","book","boot","bound","brick","broke","bruises"
,"buyers"
,"capacity","causing","cherished","chrome","classes","closely"
,"condtion","conditon","confidence","considerable","consist","consistent"
,"consumer","contents","contidion","control"
,"couldnt"
,"cream","customize","cuts"
,"daily","date","daughter"
,"deactivated","decent","deep","defender","defense","degree","demonstration"
,"depicted","depress","desciption"
,"difficulty","digi","disclaimer","discoloration","distressed","divider"
,"dlxnqat9g5wt","dock","documents","done","durable","dust","duty"
,"each","effect","either","emblem","erased","esi","essentially","etch","etched"
,"euc","every","exact","exhibition","expires"
,"facing","faded","faint","february","film","final","five","flickers"
,"folding","forgot","forwarders"
,"freezes","freight"
,"functioanlity"
,"games","generic","genuine","glitter","goes","gray","grey","guide"
,"hairline","half","handstand","hdmi","here","high","higher"
,"hold","hole","hospital"
,"immaculate","impact"
,"inivisible","instead"
,"intended","interest","interior","international","internationally"
,"into","intro"
,"ios","isnt","itself","ive"
,"jack","july"
,"keyword","kids","kind","knicks"
,"l","largest","last","late","length","let","letters","level"
,"lifting","limited","literally","literature"
,"local","logic","long","longer","looping","loose","loss","lost"
,"mars"
,"mb292ll","mc707ll","mc916ll","mc991ll","md789ll","mf432ll"
,"mic","middle","mind","minimum","mixed"
,"myself"
,"neither","newer","non","none","nonviewing","nor","november"
,"occasional","oem","often","online","oped","outside","over"
,"padfolio","pairing","paperwork","past","period","pet","piece"
,"plate","played","plug"
,"poor","portfolio","portion","pouch"
,"preinstalled","pressure","price","proof","provided"
,"ranging","rather"
,"real","realized","reassemble","reboot","receipt","recently","red"
,"reflected","refunds","relisting","remote","repeat","reponding"
,"required"
,"reserve","reshaped","residue","respond","restarts","result"
,"reviewed"
,"ringer","roughly","rubber"
,"said","same","school","scratchs","screeb"
,"seamlessly","seem","seen","semi","send","september","serious"
,"shell","shipment","short","showroom"
,"sighs","site","size"
,"sleeve","slice"
,"smoke","smooth","smudge","sn"
,"softer","software","somewhat","soon"
,"space","sparingly","sparkiling","spec","special","speck","speed","speigen"
,"stains","standup","start","status","stopped","strictly"
,"subtle","sustained"
,"swappadotcom","swiped","swivel"
,"take","technical","tempered","texture"
,"thank","then","therefore","think","those","though"
,"toddler","topback","totally","touchy","toys"
,"tried","turn","typical"
,"u","university","unknown","untouched","uppser"
,"valid","vary","viewing","virtually"
,"want","wavy","website","whole","why","winning","worn"
,"zaag","zero","zombie","zoogue"
# cor.y.train == NA
,"account","applecare","download","expect","fourth","greeting","maybe"
,"plus","purposes","significant","title","volume"
# chisq.pval high (e.g. == 1);
# keep
# carrier.fctr: "sprint", "verizon"
# cellular.fctr: "3g", "4g",wifion
# color.fctr: "gold"
# prdl.descr.my.fctr:
# storage.fctr: "128gb"
,"2016"
,"acceptable","actual","amount","awesome"
,"beautiful","before","bent","best","blocked","blocks"
,"capable","converted"
,"find"
,"gift"
,"handled","handling","headphone"
,"im","information"
,"love"
,"march","meaning","means","medium","money"
,"necessary"
,"offer","once"
,"page","product","products"
,"second","seconds","should","silver","skin","skinned"
,"tape","thoroughly","twice"
,"user"
,"way","which"
# nzv.freqRatio high (e.g. >= glb_nzv_freqCut);
# keep
# carrier.fctr: "sprint", "tmobile", "verizon"
# cellular.fctr: "3g", "4g",wifionly
# color.fctr: "gold",spacegray
# condition.fctr:
# levels:
# "Used", "For parts or not working", "Manufacturer refurbished", "New", "New other (see details)", "Seller refurbished"
# stemmed tokens:
# manufactur
# prdl.descr.my.fctr: "ipad1",ipad3,ipad4,ipadair2,ipadmini2
# storage.fctr: "128gb"
,"14","2015","3","30","4","5","6","7","8","9","90","9point5","9point7in"
,"a1432","able","about","ac","activate","activated","activation"
,"additional","adult"
,"after"
,"ago"
,"air"
,"along","already","also"
,"another","answer"
,"appears","approved","april"
,"around"
,"asis","associated"
,"auction"
,"backside","bad","battery","because","bestbuy","bezel","blue","bluetooth"
,"board","body","both","bottom","bought"
,"bright"
,"bumps","buy","buying"
,"came","camera","cameras","cannot","cant","card","carrier","cellular"
,"changed","changing","check","chip","chips"
,"color","colors","company","complete","completely","components"
,"connect","connected","connector","connects","contain","contains"
,"corporate","correctly","couple"
,"customer"
,"data"
,"dead"
,"default","defaulting","definite","definitely"
,"delivered","demo","describe","described","details"
,"do","does","dont","down"
,"drop","dropped","drops"
,"due"
,"earbuds","easily","ebay","else","engraved","engravement","engraving","entire"
,"etc"
,"even","ever","everything"
,"except","exterior","extremely"
,"fantastic","fast","faulty","features","feel","feels","fine","fix","fixed"
,"flaw","flaws"
,"frame","framing"
,"geeksquad","general","get","got","guarantee","guaranteed","guarantees"
,"hand","handful","handset","happened","hard","hardly","heavy","home","however"
,"id","if","images","imperfections"
,"inside","inspected","install","installed","instructions","invisible"
,"iphone","iphones"
,"issue","issues"
,"itunes"
,"keyboard","know","known"
,"latest","lcd","least","leather","less"
,"life","lightening","lightning","like","line","lining"
,"liquid","liquidation"
,"logo","lot","lots","lower"
,"magnetic","make","manual","manuals","many","me","memory","missing"
,"model","models","moderate","more","most","mostly"
,"multiple","musthave"
,"my"
,"name","network","networks","nice","nick","nicks"
,"nonfunctioning","noticeable","now"
,"off","old","operated","operational","otherwise","outer","own","owned"
,"party","passcode","password"
,"performance","performs","person","personal","personalized"
,"phone","physical","physically"
,"pin","pixels"
,"placed","places"
,"port","possible"
,"pretty","pristine","problem","problems","properly"
,"purchase","purchased","purchasing"
,"quality","question","questions"
,"rarely"
,"reactivated","really","rear","receive","received","regular","removed"
,"repair","retail","retina"
,"rotate","rotation"
,"running","runs"
,"s","sale","sales","sameday","scrapes","scroll","setup","several"
,"shattered","shrinkwrap","shut","shuts"
,"side","sides","sim","single"
,"so","something","sometimes","sound"
,"speaker","specifics","spent"
,"sticker","stickers","storage","store","storing","stuck","stylus"
,"super","supply","sure","surface"
,"sync"
,"taken","technician","than","these","they","thin","third","three"
,"till","tiny"
,"too","took","touch","touching","touchscreen"
,"turns"
,"two"
,"unable","under","unnoticeable","unopened","unsealed","until"
,"unused","unusable"
,"up","update","updated","updates","upgrade","upgraded","upper"
,"usage","usually"
,"verified","verify"
,"wall","water","we","week","weeks", "were","what","whats","where","while"
,"wiped","without"
,"would"
,"wrap","wrapped","wrong"
,"x"
,"year","years","your"
,"zagg","zaaginvisibleshield"
#
### !_sp
))
}## Loading required package: tm
## Loading required package: NLP
##
## Attaching package: 'NLP'
##
## The following object is masked from 'package:ggplot2':
##
## annotate#orderBy(~term, glb_post_stem_words_terms_df_lst[[txt_var]][grep("^2", glb_post_stem_words_terms_df_lst[[txt_var]]$term), ])
#glb_allobs_df[glb_post_stem_words_terms_mtrx_lst[[txt_var]][, 6] > 0, glbFeatsText]
# To identify terms with a specific freq
#paste0(sort(subset(glb_post_stop_words_terms_df_lst[[txt_var]], freq == 1)$term), collapse = ",")
#paste0(sort(subset(glb_post_stem_words_terms_df_lst[[txt_var]], freq <= 2)$term), collapse = ",")
# To identify terms with a specific freq &
# are not stemmed together later OR is value of color.fctr (e.g. gold)
#paste0(sort(subset(glb_post_stop_words_terms_df_lst[[txt_var]], (freq == 1) & !(term %in% c("blacked","blemish","blocked","blocks","buying","cables","careful","carefully","changed","changing","chargers","cleanly","cleared","connect","connects","connected","contains","cosmetics","default","defaulting","defective","definitely","describe","described","devices","displays","drop","drops","engravement","excellant","excellently","feels","fix","flawlessly","frame","framing","gentle","gold","guarantee","guarantees","handled","handling","having","install","iphone","iphones","keeped","keeps","known","lights","line","lining","liquid","liquidation","looking","lots","manuals","manufacture","minis","most","mostly","network","networks","noted","opening","operated","performance","performs","person","personalized","photograph","physically","placed","places","powering","pre","previously","products","protection","purchasing","returned","rotate","rotation","running","sales","second","seconds","shipped","shuts","sides","skin","skinned","sticker","storing","thats","theres","touching","unusable","update","updates","upgrade","weeks","wrapped","verified","verify") ))$term), collapse = ",")
#print(subset(glb_post_stem_words_terms_df_lst[[txt_var]], (freq <= 2)))
#glb_allobs_df[which(terms_mtrx[, 229] > 0), glbFeatsText]
# To identify terms with cor.y == NA
#orderBy(~-freq+term, subset(glb_post_stop_words_terms_df_lst[[txt_var]], is.na(cor.y)))
#paste(sort(subset(glb_post_stop_words_terms_df_lst[[txt_var]], is.na(cor.y))[, "term"]), collapse=",")
#orderBy(~-freq+term, subset(glb_post_stem_words_terms_df_lst[[txt_var]], is.na(cor.y)))
# To identify terms with low cor.y.abs
#head(orderBy(~cor.y.abs+freq+term, subset(glb_post_stem_words_terms_df_lst[[txt_var]], !is.na(cor.y))), 5)
# To identify terms with high chisq.pval
#subset(glb_post_stem_words_terms_df_lst[[txt_var]], chisq.pval > 0.99)
#paste0(sort(subset(glb_post_stem_words_terms_df_lst[[txt_var]], (chisq.pval > 0.99) & (freq <= 10))$term), collapse=",")
#paste0(sort(subset(glb_post_stem_words_terms_df_lst[[txt_var]], (chisq.pval > 0.9))$term), collapse=",")
#head(orderBy(~-chisq.pval+freq+term, glb_post_stem_words_terms_df_lst[[txt_var]]), 5)
#glb_allobs_df[glb_post_stem_words_terms_mtrx_lst[[txt_var]][, 68] > 0, glbFeatsText]
#orderBy(~term, glb_post_stem_words_terms_df_lst[[txt_var]][grep("^m", glb_post_stem_words_terms_df_lst[[txt_var]]$term), ])
# To identify terms with high nzv.freqRatio
#summary(glb_post_stem_words_terms_df_lst[[txt_var]]$nzv.freqRatio)
#paste0(sort(setdiff(subset(glb_post_stem_words_terms_df_lst[[txt_var]], (nzv.freqRatio >= glb_nzv_freqCut) & (freq < 10) & (chisq.pval >= 0.05))$term, c( "128gb","3g","4g","gold","ipad1","ipad3","ipad4","ipadair2","ipadmini2","manufactur","spacegray","sprint","tmobil","verizon","wifion"))), collapse=",")
# To identify obs with a txt term
#tail(orderBy(~-freq+term, glb_post_stop_words_terms_df_lst[[txt_var]]), 20)
#mydsp_obs(list(descr.my.contains="non"), cols=c("color", "carrier", "cellular", "storage"))
#grep("ever", dimnames(terms_stop_mtrx)$Terms)
#which(terms_stop_mtrx[, grep("ipad", dimnames(terms_stop_mtrx)$Terms)] > 0)
#glb_allobs_df[which(terms_stop_mtrx[, grep("16", dimnames(terms_stop_mtrx)$Terms)[1]] > 0), c(glb_category_var, "storage", txt_var)]
# To identify whether terms shd be synonyms
#orderBy(~term, glb_post_stop_words_terms_df_lst[[txt_var]][grep("^moder", glb_post_stop_words_terms_df_lst[[txt_var]]$term), ])
# term_row_df <- glb_post_stop_words_terms_df_lst[[txt_var]][grep("^came$", glb_post_stop_words_terms_df_lst[[txt_var]]$term), ]
#
# cor(glb_post_stop_words_terms_mtrx_lst[[txt_var]][glb_allobs_df$.lcn == "Fit", term_row_df$pos], glb_trnobs_df[, glb_rsp_var], use="pairwise.complete.obs")
# To identify which stopped words are "close" to a txt term
#sort(cluster_vars)
# Text Processing Step: stemDocument
# To identify stemmed txt terms
#glb_post_stop_words_terms_df_lst[[txt_var]][grep("condit", glb_post_stop_words_terms_df_lst[[txt_var]]$term), ]
#orderBy(~term, glb_post_stem_words_terms_df_lst[[txt_var]][grep("^con", glb_post_stem_words_terms_df_lst[[txt_var]]$term), ])
#glb_allobs_df[which(terms_stem_mtrx[, grep("use", dimnames(terms_stem_mtrx)$Terms)[[1]]] > 0), c(glb_id_var, "productline", txt_var)]
#glb_allobs_df[which(TfIdf_stem_mtrx[, 191] > 0), c(glb_id_var, glb_category_var, txt_var)]
#which(glb_allobs_df$UniqueID %in% c(11915, 11926, 12198))
# Text Processing Step: mycombineSynonyms
# To identify which terms are associated with not -> combine "could not" & "couldn't"
#findAssocs(glb_full_DTM_lst[[txt_var]], "not", 0.05)
# To identify which synonyms should be combined
#orderBy(~term, glb_post_stem_words_terms_df_lst[[txt_var]][grep("^c", glb_post_stem_words_terms_df_lst[[txt_var]]$term), ])
chk_comb_cor <- function(syn_lst) {
# cor(terms_stem_mtrx[glb_allobs_df$.src == "Train", grep("^(damag|dent|ding)$", dimnames(terms_stem_mtrx)[[2]])], glb_trnobs_df[, glb_rsp_var], use="pairwise.complete.obs")
print(subset(glb_post_stem_words_terms_df_lst[[txt_var]], term %in% syn_lst$syns))
print(subset(get_corpus_terms(tm_map(glb_txt_corpus_lst[[txt_var]], mycombineSynonyms, list(syn_lst), lazy=FALSE)), term == syn_lst$word))
# cor(terms_stop_mtrx[glb_allobs_df$.src == "Train", grep("^(damage|dent|ding)$", dimnames(terms_stop_mtrx)[[2]])], glb_trnobs_df[, glb_rsp_var], use="pairwise.complete.obs")
# cor(rowSums(terms_stop_mtrx[glb_allobs_df$.src == "Train", grep("^(damage|dent|ding)$", dimnames(terms_stop_mtrx)[[2]])]), glb_trnobs_df[, glb_rsp_var], use="pairwise.complete.obs")
}
#chk_comb_cor(syn_lst=list(word="cabl", syns=c("cabl", "cord")))
#chk_comb_cor(syn_lst=list(word="damag", syns=c("damag", "dent", "ding")))
#chk_comb_cor(syn_lst=list(word="dent", syns=c("dent", "ding")))
#chk_comb_cor(syn_lst=list(word="use", syns=c("use", "usag")))
glb_txt_synonyms <- list()
glb_txt_synonyms[["descr.my"]] <- #NULL : default
list(NULL
### bid._sp
# , list(word="cabl", syns=c("cabl", "cord"))#err.abs.fit.sum=26.863220
# # , list(word="charger", syns=c("charg", "charger"))
# # , list(word="come", syns=c("came", "come"))
# # , list(word="dent", syns=c("dent", "ding"))
# # , list(word="damag", syns=c(#"bad", "blemish", "broken", "crack",
# # #defect has +ve cor, others have -ve cor
# # "damag", "dent", "ding",
# # #"scratch", "scuff", "tear", "wear",
# # NULL))
# # # combining damag with defect & dent results in higher err.abs.fit.sum=26.885899
# # # combining defect with dent in higher err.abs.fit.sum=26.894976
# # , list(word="defect", syns=c(#"bad", "blemish", "broken", "crack",
# # "defect", "dent", #"ding", ding has -ve cor, others have +ve cor
# # #"scratch", "scuff", "tear", "wear",
# # NULL))
# #, list(word="new", syns=c("brand")) ???
# # , list(word="scuff", syns=c("scuf", "scuff"))
# # , list(word="show", syns=c("show", "shown"))
# # , list(word="tablet", syns=c("tab", "tablet"))
### bid._sp
### !_sp
,list(word = "cabl", syns = c("cabl", "cord"))
,list(word = "cant", syns = c("cant", "cannot"))
,list(word = "descript", syns = c("descript", "discript"))
,list(word = "generat", syns = c("gen", "generat"))
,list(word = "ipadmini", syns = c("ipadmini", "mini"))
,list(word = "kept", syns = c("keep", "kept"))
,list(word = "know", syns = c("know", "known"))
,list(word = "lightn", syns = c("lighten", "lightn"))
,list(word = "passcod", syns = c("passcod", "password"))
,list(word = "photo", syns = c("photo", "photograph", "photos", "pic", "pictur"))
,list(word = "preown", syns = c("pre", "preown", "previous",
"previouslyus", "previouslyown"))
,list(word = "protector", syns = c("protect", "protector"))
,list(word = "scuff", syns = c("scuf", "scuff"))
,list(word = "tablet", syns = c("tab", "tablet"))
,list(word = "with", syns = c("w", "with"))
,list(word = "zagg", syns = c("zagg", "zaaginvisibleshield"))
### !_sp
)
if (length(glb_txt_synonyms) > 0) names(glb_txt_synonyms) <- glbFeatsText
# Text Processing Step: filterTerms
if (!is.null(glbFeatsText)) {
require(tm)
# options include: weightTf, myweightTflog1p, myweightTfsqrt, weightTfIdf, weightBM25
glb_txt_terms_control <- list(weighting = weightTfIdf # : default
# termFreq selection criteria across obs: default: list(global=c(1, Inf))
, bounds = list(global = c(1, Inf)) # bid._sp: list(global=c(3, Inf))
# default: c(3, Inf)
, wordLengths = c(1, Inf) # bid._sp: c(2, Inf)
)
}
glb_txt_cor_var <- glb_rsp_var # bid._sp: "startprice.log10.cut.fctr" # default: glb_rsp_var
# select one from c("union.top.val.cor", "top.cor", default: "top.val", "top.chisq", "sparse")
glbFeatsTextFilter <- "top.chisq"
glbFeatsTextTermsMax <- c(15) # bid._sp: c(20) # c(50) in (old) !_sp # default: rep(10, length(glbFeatsText))
names(glbFeatsTextTermsMax) <- glbFeatsText
# Text Processing Step: extractAssoc
glbFeatsTextAssocCor <- c(1.0) #bid._sp: c(0.4) #(old) !_sp: 0.2 #default: rep(1, length(glbFeatsText))
names(glbFeatsTextAssocCor) <- glbFeatsText
# Text Processing Step: extractPatterns (ngrams)
# Potential Enhancements
# "Seller refurbished" -> D.P.refurbished.seller ?
# "Like new" -> D.P.new.like ?
# "No scratches" -> D.P.scratch.no ?
glb_important_terms <- list()
# Remember to use stemmed terms
# Have to set it even if it is not used
glb_sprs_thresholds <- c(0.950) # Generates 8 terms
# Properties:
# numrows(glb_feats_df) << numrows(glb_fitobs_df)
# Select terms that appear in at least 0.2 * O(FP/FN(glb_OOBobs_df))
# numrows(glb_OOBobs_df) = 1.1 * numrows(glb_newobs_df)
names(glb_sprs_thresholds) <- glbFeatsText
if (glb_rsp_var_raw != glb_rsp_var)
glbFeatsExclude <- union(glbFeatsExclude,
glb_rsp_var_raw)
glbFctrMaxUniqVals <- 23 # default: 20
glb_impute_na_data <- FALSE # or TRUE
glb_mice_complete.seed <- 144 # or any integer
glb_cluster <- TRUE # bid._sp:TRUE # default:FALSE
glb_cluster.seed <- 189 # or any integer
### !_sp
glb_cluster_entropy_var <- glb_rsp_var # c(glb_rsp_var, as.factor(cut(glb_rsp_var, 3)), default: NULL)
glbFeatsTextClusterVarsExclude <- FALSE # default FALSE
### bid._sp
# glb_cluster_entropy_var <- "sold" #"startprice.log10.cut.fctr"
# glbFeatsTextClusterVarsExclude <- TRUE # default FALSE
### bid._sp
glb_interaction_only_feats_lst <- list()
### bid._sp
glb_interaction_only_feats_lst[["carrier.fctr"]] <- "cellular.fctr"
glb_nzv_freqCut <- 19 # 19 is caret default
glb_nzv_uniqueCut <- 10 # 4 : bid._sp # 10 : caret::default
### bid._sp
glb_rfe_fit_sizes <- NULL
# c(106, 111, 116, 120, 128) # bid0_sp
# c(8, 11, 16, 21, 32, 64, 128) # bid1_sp
# c(47,48,49,50,51,52,53,54) # no_sp
### bid1_nosp
# rfFuncs
#c(4, 8, 16, 32, 64, 82) # Accuracy@82 = 0.8317
#c(4, 8, 16, 32, 64, 73, 82, 91) # Accuracy@82 = 0.8325
#c(16, 32, 64, 73, 76, 78, 80, 81, 82, 83) # Accuracy@82 = 0.8333
# ldaFuncs
#c(8, 16, 32, 64, 67) # Accuracy@16 = 0.8519
#c(8, 12, 16, 20, 32, 64, 67) # Accuracy@12 = 0.8519
#c(8, 10, 12, 13, 14, 15, 16, 20, 67) # Accuracy@14 = 0.8526
### bid1_nosp
### bid0_nosp
#c(4, 8, 12, 16, 20, 32, 64, 87) # Accuracy@16 = 0.7877
#c(12, 14, 15, 16, 17, 18, 20, 87) # Accuracy@17 = 0.7894
#c(8, 12, 13, 14, 15, 16, 18, 20, 87) # Accuracy@14 = 0.7900 before removing outliers
#c(8, 12, 13, 14, 15, 16, 18, 20, 87) # Accuracy@15 = 0.7868
### bid0_nosp
# outliers identified by car::outlierTest
glbObsFitOutliers <- list()
### bid._sp
# c(NULL # default: NULL
# biddable == 0 & 1; err.abs.fit.sum=423.55172
# # outliers
# , 10813 # next 665 w/ rstudent=-5.091080; biddable=3.263257; err.abs.fit.sum=418.598755
# , 10666 # next 1727 w/ rstudent=-5.163517; biddable=4.293465; err.abs.fit.sum=414.093609
# , 11736 # next 780 w/ rstudent=-5.181343; biddable=5.670483; err.abs.fit.sum=401.817992
# # old biddable importance above this
# , 10781 # next 1323 w/ rstudent=-5.151062; biddable=13.30602; err.abs.fit.sum=396.393721
# #, 10091 # next 91 w/ rstudent=-4.444452; biddable=; err.abs.fit.sum=402.673715 (up)
# #, 10166 # next 560 w/ rstudent=-5.006795; biddable=; err.abs.fit.sum=401.759324 (up)
# #, 10281 # next 281 w/ rstudent=-4.245087; biddable=; err.abs.fit.sum=401.316926 (up)
# #, 10285 # next 285 w/ rstudent=-4.483190; biddable=; err.abs.fit.sum=402.608936 (up)
# #, 10446 # next 445 w/ rstudent=-4.663418; biddable=; err.abs.fit.sum=403.074523 (up)
# #, 10542 # next 1323 w/ rstudent=-5.214517; biddable=; err.abs.fit.sum=401.04205 (up)
# #, 10543 # next 1323 w/ rstudent=-5.214517; biddable=; err.abs.fit.sum=401.04205 (up)
# #, 10561 # next 542 w/ rstudent=-4.736154; biddable=; err.abs.fit.sum=401.56198 (up)
# #, 10631 # next 166 w/ rstudent=-5.073048; biddable=; err.abs.fit.sum=401.556788 (up)
# #, 11330 # next 630 w/ rstudent=-5.117659; biddable=; err.abs.fit.sum=401.732597 (up)
# , 10091, 10166, 10281, 10285, 10446, 10542, 10543, 10561, 10631, 11330
# # biddable=18.93923; err.abs.fit.sum=359.388769
# , 10330 #biddable=19.06084; err.abs.fit.sum=355.895702
# , 10402 #biddable= 0.0 ; err.abs.fit.sum=351.315181
# , 10438 #biddable= 0.0 ; err.abs.fit.sum=347.821527
# , 10624 #biddable= 0.0 ; err.abs.fit.sum=343.724904
# , 10659 #biddable= 0.0 ; err.abs.fit.sum=331.873603
# , 11323 #biddable=10.45901; err.abs.fit.sum=324.929562
# , 11422 #biddable= 0.0 ; err.abs.fit.sum=334.839805 (up)
# biddable == 0; err.abs.fit.sum=26.713317
# , 11448 # outliers; next is 858 w/ rstudent=-5.855132; err.abs.fit.sum=24.212800
# , 11583 # outliers; next is 856 w/ rstudent=-4.792849; err.abs.fit.sum=22.164035
# , 11581 # outliers; next is 743 w/ rstudent=-4.005054; err.abs.fit.sum=18.842901
# , 10837 # outliers; next is 336 w/ rstudent=-5.279215; err.abs.fit.sum=18.124560
# , 11442 # outliers; next is 904 w/ rstudent=-4.474844; err.abs.fit.sum=15.533211
# , 11697 # outliers; next is 874 w/ rstudent=-3.678664; err.abs.fit.sum=13.829375
# , 10799 # .hatvalues == 1; total 8; iPadmini#1; err.abs.fit.sum=13.807283
# , 10017 # .hatvalues == 1; total 7; iPad3#1; err.abs.fit.sum=14.620782 (up)
# , 10027, 10859 # .hatvalues == 1; total 7; iPad1#1; err.abs.fit.sum=14.570246 (up)
# , 10332 # .hatvalues == 1; total 7; iPad4#1; err.abs.fit.sum=13.706467
# , 11759 # .hatvalues == 1; total 6; iPadAir2#1; err.abs.fit.sum=13.643043
# , 10675 # .hatvalues == 1; total 5; iPadAir#1; err.abs.fit.sum=13.623787
# , 11119 # .hatvalues == 1; total 4; iPadmini3#1; err.abs.fit.sum=NA
# , 10017, 10027, 10859 # .hatvalues == 1; total 1; iPad3#1 & iPad1#1; err.abs.fit.sum=13.438903
# biddable == 1; err.abs.fit.sum=361.78243
# , 10813 # outliers; next is 665 w/ rstudent=-5.021180; err.abs.fit.sum=356.83424
# , 10666 # outliers; next is 808 w/ rstudent=-4.764126; err.abs.fit.sum=352.46437
# , 11736 # outliers; next is 665 w/ rstudent=-4.614022; err.abs.fit.sum=348.59977
# , 10542 # outliers; next is 665 w/ rstudent=-4.654923; err.abs.fit.sum=344.18546
# , 11330 # outliers; next is 327 w/ rstudent=-4.628972; err.abs.fit.sum=336.12636
# , 10561 # outliers; next is 56 w/ rstudent=-4.612970; err.abs.fit.sum=329.50309
# , 10166 # outliers; next is 318 w/ rstudent=-4.717238; err.abs.fit.sum=318.50562
# , 10543 # outliers; next is 464 w/ rstudent=-4.811116; err.abs.fit.sum=314.32801
# , 10285 # outliers; next is 21 w/ rstudent=-4.850822; err.abs.fit.sum=310.19008
# #, 10091 # outliers; next is 464 w/ rstudent=-4.941448; err.abs.fit.sum=312.94069 (up)
# #, 10781 # outliers; next is 250 w/ rstudent=-4.793502; err.abs.fit.sum=313.03867 (up)
# , 10446 # outliers; next is 371 w/ rstudent=-4.787578; err.abs.fit.sum=307.15681
# , 10631 # outliers; next is 165 w/ rstudent=-4.130356; err.abs.fit.sum=303.34549
# #, 10330 # outliers; next is 217 w/ rstudent=-4.067684; err.abs.fit.sum=312.75121 (up)
# #, 10402 # outliers; next is 388 w/ rstudent=-4.067684; err.abs.fit.sum=311.84516 (up)
# #, 10659 # outliers; next is 128 w/ rstudent=-3.982911; err.abs.fit.sum=311.84516 (up)
# , 10091, 10781, 10330, 10402, 10659#, 10281 outliers; err.abs.fit.sum=282.381827; iPad4#0=13.806011; iPad4#1=7.799398
# #, 10281 # outliers; next is NA w/ rstudent=NA; err.abs.fit.sum=287.147331 (up); iPad4#0=14.372770; iPad4#1=4.591408
# #, 10624 # outliers; ignored along with 10281 err.abs.fit.sum=289.116467 (up); iPad4#0=; iPad4#1=
# #, 10624 # outliers; ignored w/o 10281 err.abs.fit.sum=286.415040 (up); iPad4#0=; iPad4#1=
# #, 10636 # hatvalues==1; next is 11652; err.abs.fit.sum=290.50254 (up)
# , 11652 # hatvalues==1; next is 10636; err.abs.fit.sum=282.183867
# #err.abs.fit.sum=282.227249
# )
### bid._sp
### no_sp
# # .hatvalues == 1
# c(10096, 11062, 11233, 11660, 11768)
### no_sp
### bid0_nosp
# glbObsFitOutliers[["RFE.X"]] <- c(NULL
# # is.na(RFE.X.glm$.dffits)
# , 10026, 11172, 11357, 11767
# )
# glbObsFitOutliers[["CSM.X"]] <- c(NULL
# # is.na(.dffits)
# , 10045, 10089, 10918, 11852 # iPad2#1
# , 10852, 11340, 11464 # iPadmini3#0
# , 10014, 10025, 10218, 10287, 10382, 10789, 11005, 11404, 11497, 11580, 11763, 11835 # c("Unknown#0", "iPad1#0", "iPad3#0", "iPadAir#0", "iPadAir2#0", "iPadmini#0")
# , 10155, 10362, 10746, 11029, 11451, 11724, 11759, 11799, 11807 # c("Unknown#1", "iPad1#1", "iPad2#0", "iPad3#1", "iPad4#0", "iPadAir#1", "iPadAir2#1", "iPadmini#1", "iPadmini3#1")
# , 11001 # Unknown#1
#
# # min(.rstudent) <= -121446236
# , 11760 # iPadAir2#0
# , 11751 # iPad2#1
# # max(.rstudent) >= 70875386
# , 11817 # Unknown#1
# , 10763 # iPad1#0
# , 10875 # iPad1#1
#
# # .hatvalues >= 0.99
# )
# c(NULL
# # is.na(CSM.X.glm$.dffits)
# ,10026, 10089, 10218, 10382, 10918, 11005
# # CSM.X.glm$.hatvalues > 0.99
# ,10014, 11444, 11448, 11455, 11835, 11852
# # CSM.X.glm$.rstudent <= -43969313 & >= 44111144
# ,10675, 11272
# )
### bid0_nosp
# influence.measures: car::outlier; rstudent; dffits; hatvalues; dfbeta; dfbetas
#print(outliers <- car::outlierTest(glb_models_lst[["CSM.X.glm"]]$finalModel))
#print(outliers_df <- data.frame(.Bonf.p=outliers$bonf.p))
#mdlId <- "CSM.X.glm"; mdlIdFamily <- paste0(head(unlist(str_split(mdlId, "\\.")), -1), collapse="."); obs_df <- dplyr::filter_(glb_fitobs_df, interp(~(!(var %in% glbObsFitOutliers[[mdlIdFamily]])), var = as.name(glb_id_var))); model_diags_df <- cbind(obs_df, data.frame(.rstudent=stats::rstudent(glb_models_lst[[mdlId]]$finalModel)), data.frame(.dffits=stats::dffits(glb_models_lst[[mdlId]]$finalModel)), data.frame(.hatvalues=stats::hatvalues(glb_models_lst[[mdlId]]$finalModel)));print(summary(model_diags_df[, c(".rstudent",".dffits",".hatvalues")]))
#print(subset(model_diags_df, is.na(.dffits))[, glb_id_var])
#print(model_diags_df[which.min(model_diags_df$.dffits), ])
#print(subset(model_diags_df, .hatvalues > 0.99))
#dffits_df <- merge(dffits_df, outliers_df, by="row.names", all.x=TRUE); row.names(dffits_df) <- dffits_df$Row.names; dffits_df <- subset(dffits_df, select=-Row.names)
#dffits_df <- merge(dffits_df, glb_fitobs_df, by="row.names", all.x=TRUE); row.names(dffits_df) <- dffits_df$Row.names; dffits_df <- subset(dffits_df, select=-Row.names)
#subset(dffits_df, !is.na(.Bonf.p))
#indep_vars <- myextract_actual_feats(row.names(orderBy(reformulate(c("-", paste0(glb_sel_mdl_id, ".importance"))), glb_featsimp_df))); indep_vars <- indep_vars[!grepl(".fctr", indep_vars, fixed=TRUE)]
#myplot_parcoord(obs_df=model_diags_df[, c(glb_id_var, glb_category_var, ".rstudent", ".dffits", ".hatvalues", glb_rsp_var, "startprice.log10.predict.RFE.X.glmnet", indep_vars[1:20])], obs_ix=row.names(model_diags_df) %in% names(outliers$rstudent)[1], id_var=glb_id_var, category_var=glb_category_var)
#model_diags_df[row.names(model_diags_df) %in% names(outliers$rstudent)[c(1:2)], ]
#ctgry_diags_df <- model_diags_df[model_diags_df[, glb_category_var] %in% c("Unknown#0"), ]
#myplot_parcoord(obs_df=ctgry_diags_df[, c(glb_id_var, glb_category_var, ".rstudent", ".dffits", ".hatvalues", glb_rsp_var, "startprice.log10.predict.RFE.X.glmnet", indep_vars[1:20])], obs_ix=row.names(ctgry_diags_df) %in% names(outliers$rstudent)[1], id_var=glb_id_var, category_var=glb_category_var)
#table(glb_fitobs_df[model_diags_df[, glb_category_var] %in% c("iPad1#1"), "startprice.log10.cut.fctr"])
#glb_fitobs_df[model_diags_df[, glb_category_var] %in% c("iPad1#1"), c(glb_id_var, "startprice")]
# No outliers & .dffits == NaN
#myplot_parcoord(obs_df=model_diags_df[, c(glb_id_var, glb_category_var, glb_rsp_var, "startprice.log10.predict.RFE.X.glmnet", indep_vars[1:10])], obs_ix=seq(1:nrow(model_diags_df))[is.na(model_diags_df$.dffits)], id_var=glb_id_var, category_var=glb_category_var)
#dffits_ctgry_df <- subset(dffits_df, prdl.descr.my.fctr %in% c("Unknown#0"))
#myplot_parcoord(obs_df=dffits_ctgry_df[, c(glb_id_var, glb_category_var, ".dffits", ".Bonf.p", glb_rsp_var, "startprice.log10.predict.RFE.X.glmnet", indep_vars[1:5])], obs_ix=seq(1:nrow(dffits_ctgry_df))[!is.na(dffits_ctgry_df$.Bonf.p)], id_var=glb_id_var, category_var=glb_category_var)
#
#car::influenceIndexPlot(glb_models_lst[["RFE.X.glm"]]$finalModel, id.n=3)
# myplot_parcoord(obs_df=glb_fitobs_df[, c(glb_id_var, glb_rsp_var,
# "startprice.log10.predict.RFE.X.glmnet",
# indep_vars[1:5])], obs_ix=hatobs_ix, id_var=glb_id_var)
# myplot_parcoord(x=glb_fitobs_df[, c(glb_rsp_var, "startprice.log10.predict.RFE.X.glmnet",
# indep_vars[1:2])], obs_ix=hatobs_ix)
# hatvals <- hatvalues(glb_models_lst[["RFE.X.glm"]]$finalModel)
# hatobs_ix <- which(hatvals == max(hatvals))
# MASS::parcoord(x=glb_fitobs_df[, c(glb_rsp_var, "startprice.log10.predict.RFE.X.glmnet",
# indep_vars[1:2])], var.label=TRUE)
#plot(hatvalues(glb_models_lst[["RFE.X.glm"]]$finalModel), type = "h")
#glb_fitobs_df[which(row.names(glb_fitobs_df) %in% c("972")), c(glb_id_var, glb_rsp_var, glb_rsp_var_raw, "sold", glb_category_var)]
#all.equal(glb_models_lst[[glb_sel_mdl_id]], glb_models_lst[[glb_fin_mdl_id]])
glbObsTrnOutliers <- list()
#car::outlierTest(glb_models_lst[["RFE.X.glm"]]$finalModel)
#glb_trnobs_df[which(row.names(glb_fitobs_df) %in% c("972")), c(glb_id_var, glb_rsp_var, glb_rsp_var_raw, "sold", glb_category_var)]
glb_models_lst <- list(); glb_models_df <- data.frame()
# Regression
if (glb_is_regression) {
glbMdlMethods <- c(NULL
# deterministic
#, "lm",
, "glm"
#, "bayesglm" # crashing w/ parallel processing
, "glmnet", "rpart"
# non-deterministic
, "gbm", "rf"
# Unknown
#, "nnet" , "avNNet" # predicts 1 for all obs in bid0_sp # runs 25 models per cv sample for tunelength=5
, "svmLinear", "svmLinear2"
#, "svmPoly" # crashing w/ parallel processing #, "svmPoly" runs 75 models per cv sample for tunelength=5
#, "svmRadial" # crashing w/ parallel processing
, "earth"
#, "bagEarth" # Takes a long time
)
} else
# Classification - Add ada,bagEarth (auto feature selection)
if (glb_is_binomial)
glbMdlMethods <- c(NULL
# deterministic
, "glm"
, "bayesglm" # crashing w/ parallel processing
, "glmnet"
, "nnet"
, "rpart"
# non-deterministic
, "gbm"
, "avNNet" # runs 25 models per cv sample for tunelength=5
, "rf"
# Unknown
, "lda", "lda2"
# svm models crash when predict is called -> internal to kernlab it should call predict without .outcome
# , "svmLinear", "svmLinear2", "svmPoly" #, "svmPoly" runs 75 models per cv sample for tunelength=5
# , "svmRadial"
, "earth"
, "bagEarth" # Takes a long time
) else
glbMdlMethods <- c("rpart", "rf", "gbm")
glb_mdl_family_lst <- list(); glb_mdl_feats_lst <- list()
# family: Choose from c("RFE.X", "CSM.X", "All.X", "Best.Interact")
# methods: Choose from c(NULL, <method>, glbMdlMethods)
glb_mdl_family_lst[["RFE.X"]] <- c("glmnet", "glm") # # setdiff(glbMdlMethods, c("lda", "lda2")) # crashing due to category:.clusterid #non-NULL list is mandatory
glb_mdl_family_lst[["All.X"]] <- "glmnet" # non-NULL list is mandatory
glb_mdl_family_lst[["Best.Interact"]] <- "glmnet" # non-NULL list is mandatory
### bid1_sp
# glb_mdl_family_lst[["CSM.X"]] <- "glmnet"
# glb_mdl_feats_lst[["CSM.X"]] <- c(NULL
# # from RFE.X
# , "startprice.dgt1.is9", "startprice.dcm2.is9", "startprice.dcm1.is9", "startprice.dgt2.is9"
# #, "condition.fctr"
# , "prdl.descr.my.fctr", "color.fctr"
# #, "D.ratio.weight.sum.wrds.n"
# , "cellular.fctr", "cellular.fctr:carrier.fctr"
#
# # from RFE.X.Interact
# , "cellular.fctr:prdl.descr.my.fctr", "cellular.fctr:startprice.dgt2.is9", "cellular.fctr:startprice.dgt1.is9", "cellular.fctr:color.fctr"
# , "cellular.fctr:condition.fctr" # RMSE up with keeping condition.fctr in the model
# # RMSE & R.sq up with removing condition.fctr from the model
# , "cellular.fctr:D.ratio.weight.sum.wrds.n"
# )
### bid1_sp
### !_sp
# glb_mdl_family_lst[["CSM.X"]] <- "glmnet"
# glb_mdl_feats_lst[["CSM.X"]] <- c(NULL
# , "D.ratio.sum.TfIdf.wrds.n", "D.TfIdf.sum.stem.stop.Ratio",
# , "D.npnct15.log", "D.npnct03.log", "D.wrds.n.log", "D.chrs.n.log")
# indep_vars <- union(setdiff(indep_vars, interact_vars_vctr),
# paste(glb_category_var, interact_vars_vctr,
# sep=ifelse(grepl("\\.fctr", glb_category_var), "*", ".fctr*")))
# indep_vars <- union(setdiff(indep_vars,
# c("startprice.log.diff", "startprice.unit9", "biddable", "cellular.fctr", "carrier.fctr")),
# c("startprice.log.diff*biddable", "startprice.unit9*biddable", "cellular.fctr*carrier.fctr"))
### !_sp
# Check if interaction features make fit better
# Check if tuning parameters make fit better
glb_tune_models_df <- data.frame()
#RFE.X.avNNet
### bid0_sp
# size=[1] 3 5 7 9; decay=[0] 1e-04 0.001 0.01 0.1; bag=[FALSE]; RMSE=1.3300906
### bid1_sp
# size=[1] 3 5 7 9; decay=0 0.0001 [0.001] 0.01 0.1; bag=[FALSE]; RMSE=0.9285472
### bid0&1_sp
#RFE.X.bagEarth
### bid0_sp
#RFE.X.bagEarth degree=[1]; nprune=[33]; RMSE=0.1507259
### bid1_sp
#RFE.X.bagEarth degree=[1]; nprune=[32]; RMSE=0.6379639
#RFE.X.bagEarth degree=[1] 2 3; nprune=8 16 32 64 [128]; RMSE=0.6334405
#RFE.X.bagEarth degree=1 [2]; nprune=16 32 64 128 [256]; RMSE=0.6211924
#RFE.X.bagEarth degree=1 [2]; nprune=64 128 200 225 [256]; RMSE=0.6320776 (up)
#RFE.X.bagEarth degree=[1] 2; nprune=64 128 225 256 [275]; RMSE=0.640644 (up)
#RFE.X.bagEarth degree=1 [2] 3; nprune=64 128 200 [256] 300; RMSE=0.6496039 (up)
#RFE.X.bagEarth degree=1 [2] 3; nprune=32 64 128 256 [512]; RMSE=0.6404529 (up)
#RFE.X.bagEarth degree=1 [2] 3; nprune=64 128 256 512 [1024]; RMSE=0.6486663 (up)
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method = "bagEarth", parameter = "nprune", vals = "256")
# ,data.frame(method = "bagEarth", parameter = "degree", vals = "2")
# ))
### bid0&1_sp
### bid0_sp
#RFE.X.earth degree=[1]; nprune=2 [9] 17 25 33; RMSE=0.1334478
### bid0_sp
#RFE.X.gbm
### bid0_sp
# shrinkage=[0.1]; n.trees=50 100 150 [200] 250; RMSE=0.2062651
# shrinkage=0.00 0.05 0.10 0.15 [0.20]; n.trees=50 [100] 150 200 250; interaction.depth=1 [2] 3 4 5; n.minobsinnode=[10]; RMSE=0.2019453
# shrinkage=0.00 0.05 [0.10] 0.15 0.20; n.trees=50 100 150 200 [250]; interaction.depth=[1] 2 3 4 5; n.minobsinnode=[10]; RMSE=0.2008313
# shrinkage=0.05 [0.10] 0.15 0.20 0.25; n.trees=100 150 200 [250] 300; interaction.depth=[1] 2 3 4 5; n.minobsinnode=[10]; RMSE=0.2008313
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method = "gbm", parameter = "shrinkage", min = 0.05, max = 0.25, by = 0.05)
# ,data.frame(method = "gbm", parameter = "n.trees", min = 100, max = 300, by = 50)
# ,data.frame(method = "gbm", parameter = "interaction.depth", min = 1, max = 5, by = 1)
# ,data.frame(method = "gbm", parameter = "n.minobsinnode", min = 10, max = 10, by = 10)
# #seq(from=0.05, to=0.25, by=0.05)
# ))
### bid1_sp
# shrinkage=[0.1]; n.trees=50 100 150 200 [250]; interaction.depth=1 2 3 4 [5]; n.minobsinnode=[10]; RMSE=0.5054172
# shrinkage=0.03 [0.04] 0.05 0.06 0.07; n.trees=100 [150] 200 250 300; interaction.depth=2 3 4 5 [6]; n.minobsinnode=6 [8] 10 12 14; RMSE=0.5036430
# shrinkage=0.03 [0.04] 0.05 0.06 0.07; n.trees=100 150 [200] 250 300; interaction.depth=3 4 5 [6] 7; n.minobsinnode=6 8 [10] 12 14; RMSE=0.502774
# shrinkage=0.04; n.trees=200; interaction.depth=6; n.minobsinnode=10; RMSE=0.502774
# shrinkage=[0.05] 0.10 0.15 0.20 0.25; n.trees=100 [150] 200 250 300; interaction.depth=2 3 [4] 5 6; n.minobsinnode=[10]; RMSE=0.5058678 (up)
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method = "gbm", parameter = "shrinkage", vals = "0.04")
# ,data.frame(method = "gbm", parameter = "n.trees", vals = "200")
# ,data.frame(method = "gbm", parameter = "interaction.depth", vals = "6")
# ,data.frame(method = "gbm", parameter = "n.minobsinnode", vals = "10")
# ))
### bid0&1_sp
#RFE.X.glmnet
### bid1_sp
# alpha=0.100 [0.325] 0.550 0.775 1.000; lambda=0.0005232693 0.0024288010 0.0112734954 [0.0523269304] 0.2428800957; RMSE=0.6164891
### bid1_nosp
# alpha=0.325 0.550 0.775 0.8875 [1.000]; lambda=9.858855e-05 [0.0001971771] 0.0009152152 0.0042480525 0.0197177130; Accuracy=0.8455510
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method = "glmnet", parameter = "alpha", vals = "0.550 0.775 0.8875 0.94375 1.000")
# ,data.frame(method = "glmnet", parameter = "lambda", vals = "9.858855e-05 0.0001971771 0.0009152152 0.0042480525 0.0197177130")
# ))
#RFE.X.nnet
### bid0_sp
# size=[1] 3 5 7 9; decay=[0] 1e-04 0.001 0.01 0.1; RMSE=1.3300906
### bid1_sp
# size=1 3 5 7 [9]; decay=0e+00 1e-04 1e-03 1e-02 [1e-01]; RMSE=0.9289109
# size=3 5 [7] 9 11; decay=0.0001 0.001 0.01 [0.1] 0.2; RMSE=0.9287422
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method = "nnet", parameter = "size", vals = "3 5 7 9 11")
# ,data.frame(method = "nnet", parameter = "decay", vals = "0.0001 0.0010 0.0100 0.1000 0.2000")
# ))
### bid0&1_sp
#RFE.X.rf # Don't bother; results are not deterministic
### bid0_sp
# mtry=2 35 [68] 101 134; RMSE=0.1331992
# mtry=2 35 68 [101] 134; RMSE=0.1339974
### bid0_sp
### no_sp
# mtry=2 41 81 121 [161]; Accuracy=0.8398314
# mtry=41 81 [121] 161 181; Accuracy=0.8282403
### no_sp
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method = "rf", parameter = "mtry", vals = "2 5 9 13 17")
# ))
#RFE.X.rpart
### bid0_sp
# cp=0.020 [0.025] 0.030 0.035 0.040; RMSE=0.1770237
### bid1_sp
# cp=0.001 [0.003] 0.005 0.007 0.009; RMSE=0.5186586
### bid0_nosp
# cp=0.004347826 [0.008695652] 0.017391304 0.021739130 0.034782609; Accuracy=0.8120340
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method = "rpart", parameter = "cp", vals = "0.004347826 0.008695652 0.017391304 0.021739130 0.034782609")
# ))
#RFE.X.svmLinear
### bid0_sp
# C=[1]; RMSE=0.1374094
# C=1e-02 [0.1] 5e-01 1e+00 2e+00 3e+00 4e+00 1e+01 1e+02; RMSE=0.1271318
# C=0.01 0.05 [0.10] 0.50 1.00 2.00 3.00 4.00; RMSE=0.1271318; 0.1296718
### bid1_sp
# C=[1]; RMSE=0.6614060
# C=1e-02 [1e-01] 1e+00 1e+01 1e+02; RMSE=0.6373977
# C=[0.05] 0.10 0.50 1.00 10.00; RMSE=0.6324697
# C=0.01 [0.05] 0.10 0.50 1.00; RMSE=0.6324697
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method = "svmLinear", parameter = "C", vals = "0.01 0.05 0.1 0.5 1")
# ))
### bid0&1_sp
#RFE.X.svmLinear2
### bid0_sp
# cost=[0.25] 0.50 1.00 2.00 4.00; RMSE=0.1276354
# cost=0.0625 0.1250 [0.25] 0.50 1.00; RMSE=0.1276354
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method = "svmLinear2", parameter = "cost", vals = "0.0625 0.125 0.25 0.5 1")
### bid1_sp
# cost=[0.25] 0.50 1.00 2.00 4.00; RMSE=0.6483622
# cost=[0.0625] 0.1250 0.25 0.50 1.00; RMSE=0.6335311
# cost=0.0312 [0.0625] 0.1250 0.25 0.50; RMSE=0.6335311
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method = "svmLinear2", parameter = "cost", vals = "0.0312 0.0625 0.125 0.25 0.50")
# ))
### bid0&1_sp
#RFE.X.svmPoly
### bid0_sp
# degree=[1] 2 3; scale=0.001 0.01 [0.1] 1 10; C=0.25 0.50 1.00 [2.00] 4.00; RMSE=0.1276130
# degree=[1] 2 3 4 5; scale=0.01 0.05 [0.1] 0.5 1; C=0.50 1.00 [2.00] 3.00 4.00; RMSE=0.1276130
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method="svmPoly", parameter="degree", min=1, max=5, by=1) #seq(1, 5, 1)
# ))
# glb_tune_models_df <- myrbind_df(glb_tune_models_df, rbind(data.frame()
# ,data.frame(method="svmPoly", parameter="scale", vals="0.01, 0.05, 0.1, 0.5, 1")
# ,data.frame(method="svmPoly", parameter="C", vals="0.50, 1.00, 2.00, 3.00, 4.00")
# ))
### bid0_sp
#RFE.X.svmRadial
### bid0_sp
# sigma=[0.08674323]; C=0.25 0.50 1.00 [2.00] 4.00; RMSE=0.1614957
### bid0_sp
#data.frame(parameter="mtry", min=080, max=100, by=10),
#glb_to_sav(); all.equal(sav_models_df, glb_models_df)
#glb_models_df <- subset(sav_models_df, id != "RFE.X.gbm"); print(sort(glb_models_df$id))
glb_preproc_methods <- NULL
### bid0_sp
# c("YeoJohnson", "center.scale",
# # crashes with train: all the RMSE metric values are missing
# # probably due to interaction vars
# "range", "pca", "ica",
# "spatialSign")
### bid0_sp
### bid1_sp
# c("YeoJohnson", "center.scale", "range", "pca", "ica", "spatialSign")
### bid1_sp
# Baseline prediction model feature(s)
glb_Baseline_mdl_var <- NULL # or c("<col_name>")
glbMdlMetric_terms <- NULL # or matrix(c(
# 0,1,2,3,4,
# 2,0,1,2,3,
# 4,2,0,1,2,
# 6,4,2,0,1,
# 8,6,4,2,0
# ), byrow=TRUE, nrow=5)
glbMdlMetricSummary <- NULL # or "<metric_name>"
glbMdlMetricMaximize <- NULL # or FALSE (TRUE is not the default for both classification & regression)
glbMdlMetricSummaryFn <- NULL # or function(data, lev=NULL, model=NULL) {
# confusion_mtrx <- t(as.matrix(confusionMatrix(data$pred, data$obs)))
# #print(confusion_mtrx)
# #print(confusion_mtrx * glbMdlMetric_terms)
# metric <- sum(confusion_mtrx * glbMdlMetric_terms) / nrow(data)
# names(metric) <- glbMdlMetricSummary
# return(metric)
# }
glb_rcv_n_folds <- 3 # or NULL
glb_rcv_n_repeats <- 3 # or NULL
glb_clf_proba_threshold <- NULL # 0.5
# Model selection criteria
if (glb_is_regression)
glbMdlMetricsEval <- c("min.RMSE.OOB", "max.R.sq.OOB", "max.Adj.R.sq.fit")
#glbMdlMetricsEval <- c("min.RMSE.fit", "max.R.sq.fit", "max.Adj.R.sq.fit")
if (glb_is_classification) {
if (glb_is_binomial)
glbMdlMetricsEval <-
c("max.Accuracy.OOB", "max.AUCROCR.OOB", "max.AUCpROC.OOB", "min.aic.fit", "max.Accuracy.fit") else
glbMdlMetricsEval <- c("max.Accuracy.OOB", "max.Kappa.OOB")
}
# select from NULL [no ensemble models], "auto" [all models better than MFO or Baseline], c(mdl_ids in glb_models_lst) [Typically top-rated models in auto]
glb_mdl_ensemble <- NULL
### bid0_sp
# c("RFE.X.glm"
# #, "RFE.X.bayesglm"
# , "RFE.X.glmnet", "RFE.X.rpart", "RFE.X.gbm", "RFE.X.rf", "RFE.X.svmLinear", "RFE.X.svmLinear2"
# #, "RFE.X.svmPoly", "RFE.X.svmRadial"
# , "RFE.X.earth", "RFE.X.bagEarth", "RFE.X.Interact.glmnet", "RFE.X.YeoJohnson.glmnet", "RFE.X.center.scale.glmnet", "RFE.X.spatialSign.glmnet")
### bid1_sp
# "auto"; err.abs.fit.sum=76.699774; min.RMSE.fit=0.2186429
# "RFE.X.*"; err.abs.fit.sum=; min.RMSE.fit=0.221114
# c("RFE.X.spatialSign.rf", "RFE.X.YeoJohnson.rf", "RFE.X.center.scale.rf", "RFE.X.rf", "RFE.X.avNNet", "RFE.X.bagEarth", "RFE.X.earth", "RFE.X.gbm", "RFE.X.glmnet", "RFE.X.nnet", "RFE.X.svmLinear2", "RFE.X.glm", "RFE.X.svmLinear", "RFE.X.rpart")
### bid1_sp
### no_sp
# c("RFE.X.rpart", "RFE.X.earth", "RFE.X.glmnet", "RFE.X.glm", "RFE.X.bayesglm", "RFE.X.nnet", "RFE.X.avNNet", "RFE.X.Interact.glmnet", "RFE.X.gbm"
# #,"RFE.X.bagEarth" #takes a long time
# )
### no_sp
### bid1_nosp
# c("CSM.X.glm", "CSM.X.bayesglm", "CSM.X.glmnet", "CSM.X.nnet", "CSM.X.avNNet", "CSM.X.rpart", "CSM.X.gbm", "CSM.X.rf", "CSM.X.lda", "CSM.X.lda2", "CSM.X.earth", "CSM.X.bagEarth", "CSM.X.Interact.glmnet") # Accuracy.OOB(Ensemble.repeatedcv.glmnet) = 0.8240223
# Deleted models with repeated cor(CSM.X.gbm.prob)
# c("CSM.X.glm", "CSM.X.bayesglm", "CSM.X.glmnet", "CSM.X.nnet", "CSM.X.avNNet", "CSM.X.rpart", "CSM.X.gbm", "CSM.X.rf", "CSM.X.lda2", "CSM.X.earth", "CSM.X.bagEarth") # Accuracy.OOB(Ensemble.repeatedcv.glmnet) = 0.8240223
# Used step(ntv.glm) to select simplest model
# c("CSM.X.glmnet", "CSM.X.nnet", "CSM.X.avNNet", "CSM.X.rpart", "CSM.X.gbm", "CSM.X.rf", "CSM.X.earth", "CSM.X.bagEarth") # Accuracy.OOB(Ensemble.repeatedcv.glmnet) = 0.8156425
# Deleted models with cor(CSM.X.gbm.prob) between Q3 & Max
# c("CSM.X.glm", "CSM.X.bayesglm", "CSM.X.glmnet", "CSM.X.nnet", "CSM.X.avNNet", "CSM.X.rpart", "CSM.X.gbm", "CSM.X.rf", "CSM.X.lda2") # Accuracy.OOB(Ensemble.repeatedcv.glmnet) = 0.8184358
### bid1_nosp
#paste(grep("CSM.X", names(glb_models_lst), fixed=TRUE, value=TRUE), collapse='", "')
# Only for classifications; for regressions remove "(.*)\\.prob" form the regex
# tmp_fitobs_df <- glb_fitobs_df[, grep(paste0("^", gsub(".", "\\.", glb_rsp_var_out, fixed = TRUE), "CSM\\.X\\.(.*)\\.prob"), names(glb_fitobs_df), value = TRUE)]; cor_mtrx <- cor(tmp_fitobs_df); cor_vctr <- sort(cor_mtrx[row.names(orderBy(~-Overall, varImp(glb_models_lst[["Ensemble.repeatedcv.glmnet"]])$importance))[1], ]); summary(cor_vctr); cor_vctr
#ntv.glm <- glm(reformulate(indep_vars, glb_rsp_var), family = "binomial", data = glb_fitobs_df)
#step.glm <- step(ntv.glm)
glb_sel_mdl_id <- "All.X.glmnet" #NULL #select from c(NULL, "All.X.glmnet", "RFE.X.glmnet")
glb_fin_mdl_id <- NULL #select from c(NULL, glb_sel_mdl_id)
glb_dsp_cols <- c("sold", ".grpid", "color", "condition", "cellular", "carrier", "storage", "biddable", "startprice", "sprice.predict.diff")
glb_out_obs <- NULL # "all" for bid._sp # select from c(NULL, "all", "new", "trn")
glb_out_vars_lst <- list()
# glb_id_var will be the first output column, by default
### !_sp
glb_out_vars_lst[["Probability1"]] <- "%<d-% paste0(glb_rsp_var_out, glb_fin_mdl_id, '.prob')"
### !_sp
### bid._sp
# glb_out_vars_lst[[glb_rsp_var_raw]] <- glb_rsp_var_raw
# glb_out_vars_lst[[paste0(head(unlist(strsplit(glb_rsp_var_out, "")), -1), collapse = "")]] <-
# "%<d-% paste0(glb_rsp_var_out, glb_fin_mdl_id)"
### bid._sp
glbOutStackFnames <- NULL #: default
# c("ebayipads_txt_assoc1_out_bid1_stack.csv") # manual stack
# c("ebayipads_finmdl_bid1_out_nnet_1.csv") # universal stack
glb_out_pfx <- "ebayipads_txt2_n15_"
glb_save_envir <- FALSE # or TRUE
# Depict process
glb_analytics_pn <- petrinet(name = "glb_analytics_pn",
trans_df = data.frame(id = 1:6,
name = c("data.training.all","data.new",
"model.selected","model.final",
"data.training.all.prediction","data.new.prediction"),
x=c( -5,-5,-15,-25,-25,-35),
y=c( -5, 5, 0, 0, -5, 5)
),
places_df=data.frame(id=1:4,
name=c("bgn","fit.data.training.all","predict.data.new","end"),
x=c( -0, -20, -30, -40),
y=c( 0, 0, 0, 0),
M0=c( 3, 0, 0, 0)
),
arcs_df=data.frame(
begin=c("bgn","bgn","bgn",
"data.training.all","model.selected","fit.data.training.all",
"fit.data.training.all","model.final",
"data.new","predict.data.new",
"data.training.all.prediction","data.new.prediction"),
end =c("data.training.all","data.new","model.selected",
"fit.data.training.all","fit.data.training.all","model.final",
"data.training.all.prediction","predict.data.new",
"predict.data.new","data.new.prediction",
"end","end")
))
#print(ggplot.petrinet(glb_analytics_pn))
print(ggplot.petrinet(glb_analytics_pn) + coord_flip())## Loading required package: grid
glb_analytics_avl_objs <- NULL
glb_chunks_df <- myadd_chunk(NULL, "import.data")## label step_major step_minor label_minor bgn end elapsed
## 1 import.data 1 0 0 23.954 NA NAStep 1.0: import data
chunk option: eval=
#glb_chunks_df <- myadd_chunk(NULL, "import.data")
glb_to_sav <- function() {
sav_allobs_df <<- glb_allobs_df
sav_trnobs_df <<- glb_trnobs_df
if (any(grepl("glb_fitobs_df", ls(envir=globalenv()), fixed=TRUE)) &&
!is.null(glb_fitobs_df)) sav_fitobs_df <<- glb_fitobs_df
if (any(grepl("glb_OOBobs_df", ls(envir=globalenv()), fixed=TRUE)) &&
!is.null(glb_OOBobs_df)) sav_OOBobs_df <<- glb_OOBobs_df
if (any(grepl("glb_newobs_df", ls(envir=globalenv()), fixed=TRUE)) &&
!is.null(glb_newobs_df)) {
#print("Attempting to save glb_newobs_df...")
sav_newobs_df <<- glb_newobs_df
}
if (any(grepl("glb_ctgry_df", ls(envir=globalenv()), fixed=TRUE)) &&
!is.null(glb_ctgry_df)) sav_ctgry_df <<- glb_ctgry_df
if (!is.null(glb_models_lst )) sav_models_lst <<- glb_models_lst
if (!is.null(glb_models_df )) sav_models_df <<- glb_models_df
if (any(grepl("glb_feats_df", ls(envir=globalenv()), fixed=TRUE)) &&
!is.null(glb_feats_df)) sav_feats_df <<- glb_feats_df
if (any(grepl("glb_featsimp_df", ls(envir=globalenv()), fixed=TRUE)) &&
!is.null(glb_featsimp_df)) sav_featsimp_df <<- glb_featsimp_df
}
glb_trnobs_df <- myimport_data(url=glb_trnng_url, comment="glb_trnobs_df",
force_header=TRUE)## [1] "Reading file ./data/eBayiPadTrain.csv..."
## [1] "dimensions of data in ./data/eBayiPadTrain.csv: 1,861 rows x 11 cols"
## description
## 1 iPad is in 8.5+ out of 10 cosmetic condition!
## 2 Previously used, please read description. May show signs of use such as scratches to the screen and
## 3
## 4
## 5 Please feel free to buy. All products have been thoroughly inspected, cleaned and tested to be 100%
## 6
## biddable startprice condition cellular carrier color
## 1 0 159.99 Used 0 None Black
## 2 1 0.99 Used 1 Verizon Unknown
## 3 0 199.99 Used 0 None White
## 4 0 235.00 New other (see details) 0 None Unknown
## 5 0 199.99 Seller refurbished Unknown Unknown Unknown
## 6 1 175.00 Used 1 AT&T Space Gray
## storage productline sold UniqueID
## 1 16 iPad 2 0 10001
## 2 16 iPad 2 1 10002
## 3 16 iPad 4 1 10003
## 4 16 iPad mini 2 0 10004
## 5 Unknown Unknown 0 10005
## 6 32 iPad mini 2 1 10006
## description
## 65
## 283 Pristine condition, comes with a case and stylus.
## 948 \211\333\317Used Apple Ipad 16 gig 1st generation in Great working condition and 100% functional.Very little
## 1354
## 1366 Item still in complete working order, minor scratches, normal wear and tear but no damage. screen is
## 1840
## biddable startprice condition cellular carrier color
## 65 0 195.00 Used 0 None Unknown
## 283 1 20.00 Used 0 None Unknown
## 948 0 110.00 Seller refurbished 0 None Black
## 1354 0 300.00 Used 0 None White
## 1366 1 125.00 Used Unknown Unknown Unknown
## 1840 0 249.99 Used 1 Sprint Space Gray
## storage productline sold UniqueID
## 65 16 iPad mini 0 10065
## 283 64 iPad 1 0 10283
## 948 32 iPad 1 0 10948
## 1354 16 iPad Air 1 11354
## 1366 Unknown iPad 1 1 11366
## 1840 16 iPad Air 1 11840
## description
## 1856 Overall item is in good condition and is fully operational and ready to use. Comes with box and
## 1857 Used. Tested. Guaranteed to work. Physical condition grade B+ does have some light scratches and
## 1858 This item is brand new and was never used; however, the box and/or packaging has been opened.
## 1859
## 1860 This unit has minor scratches on case and several small scratches on the display. \nIt is in
## 1861 30 Day Warranty. Fully functional engraved iPad 1st Generation with signs of normal wear which
## biddable startprice condition cellular carrier
## 1856 0 89.50 Used 1 AT&T
## 1857 0 239.95 Used 0 None
## 1858 0 329.99 New other (see details) 0 None
## 1859 0 400.00 New 0 None
## 1860 0 89.00 Seller refurbished 0 None
## 1861 0 119.99 Used 1 AT&T
## color storage productline sold UniqueID
## 1856 Unknown 16 iPad 1 0 11856
## 1857 Black 32 iPad 4 1 11857
## 1858 Space Gray 16 iPad Air 0 11858
## 1859 Gold 16 iPad mini 3 0 11859
## 1860 Black 64 iPad 1 1 11860
## 1861 Black 64 iPad 1 0 11861
## 'data.frame': 1861 obs. of 11 variables:
## $ description: chr "iPad is in 8.5+ out of 10 cosmetic condition!" "Previously used, please read description. May show signs of use such as scratches to the screen and " "" "" ...
## $ biddable : int 0 1 0 0 0 1 1 0 1 1 ...
## $ startprice : num 159.99 0.99 199.99 235 199.99 ...
## $ condition : chr "Used" "Used" "Used" "New other (see details)" ...
## $ cellular : chr "0" "1" "0" "0" ...
## $ carrier : chr "None" "Verizon" "None" "None" ...
## $ color : chr "Black" "Unknown" "White" "Unknown" ...
## $ storage : chr "16" "16" "16" "16" ...
## $ productline: chr "iPad 2" "iPad 2" "iPad 4" "iPad mini 2" ...
## $ sold : int 0 1 1 0 0 1 1 0 1 1 ...
## $ UniqueID : int 10001 10002 10003 10004 10005 10006 10007 10008 10009 10010 ...
## - attr(*, "comment")= chr "glb_trnobs_df"
## NULL# glb_trnobs_df <- read.delim("data/hygiene.txt", header=TRUE, fill=TRUE, sep="\t",
# fileEncoding='iso-8859-1')
# glb_trnobs_df <- read.table("data/hygiene.dat.labels", col.names=c("dirty"),
# na.strings="[none]")
# glb_trnobs_df$review <- readLines("data/hygiene.dat", n =-1)
# comment(glb_trnobs_df) <- "glb_trnobs_df"
# glb_trnobs_df <- data.frame()
# for (symbol in c("Boeing", "CocaCola", "GE", "IBM", "ProcterGamble")) {
# sym_trnobs_df <-
# myimport_data(url=gsub("IBM", symbol, glb_trnng_url), comment="glb_trnobs_df",
# force_header=TRUE)
# sym_trnobs_df$Symbol <- symbol
# glb_trnobs_df <- myrbind_df(glb_trnobs_df, sym_trnobs_df)
# }
# glb_trnobs_df <-
# glb_trnobs_df %>% dplyr::filter(Year >= 1999)
if (glb_is_separate_newobs_dataset) {
glb_newobs_df <- myimport_data(url=glb_newdt_url, comment="glb_newobs_df",
force_header=TRUE)
# To make plots / stats / checks easier in chunk:inspectORexplore.data
glb_allobs_df <- myrbind_df(glb_trnobs_df, glb_newobs_df);
comment(glb_allobs_df) <- "glb_allobs_df"
} else {
glb_allobs_df <- glb_trnobs_df; comment(glb_allobs_df) <- "glb_allobs_df"
if (!glb_split_entity_newobs_datasets) {
stop("Not implemented yet")
glb_newobs_df <- glb_trnobs_df[sample(1:nrow(glb_trnobs_df),
max(2, nrow(glb_trnobs_df) / 1000)),]
} else if (glb_split_newdata_method == "condition") {
glb_newobs_df <- do.call("subset",
list(glb_trnobs_df, parse(text=glb_split_newdata_condition)))
glb_trnobs_df <- do.call("subset",
list(glb_trnobs_df, parse(text=paste0("!(",
glb_split_newdata_condition,
")"))))
} else if (glb_split_newdata_method == "sample") {
require(caTools)
set.seed(glb_split_sample.seed)
split <- sample.split(glb_trnobs_df[, glb_rsp_var_raw],
SplitRatio=(1-glb_split_newdata_size_ratio))
glb_newobs_df <- glb_trnobs_df[!split, ]
glb_trnobs_df <- glb_trnobs_df[split ,]
} else if (glb_split_newdata_method == "copy") {
glb_trnobs_df <- glb_allobs_df
comment(glb_trnobs_df) <- "glb_trnobs_df"
glb_newobs_df <- glb_allobs_df
comment(glb_newobs_df) <- "glb_newobs_df"
} else stop("glb_split_newdata_method should be %in% c('condition', 'sample', 'copy')")
comment(glb_newobs_df) <- "glb_newobs_df"
myprint_df(glb_newobs_df)
str(glb_newobs_df)
if (glb_split_entity_newobs_datasets) {
myprint_df(glb_trnobs_df)
str(glb_trnobs_df)
}
} ## [1] "Reading file ./data/eBayiPadTest.csv..."
## [1] "dimensions of data in ./data/eBayiPadTest.csv: 798 rows x 10 cols"
## description
## 1 like new
## 2 Item is in great shape. I upgraded to the iPad Air 2 and don't need the mini any longer, even though
## 3 This iPad is working and is tested 100%. It runs great. It is in good condition. Cracked digitizer.
## 4
## 5 Grade A condition means that the Ipad is 100% working condition. Cosmetically 8/9 out of 10 - Will
## 6 Brand new factory sealed iPad in an OPEN BOX...THE BOX ITSELF IS HEAVILY DISTRESSED(see
## biddable startprice condition cellular carrier color
## 1 0 105.00 Used 1 AT&T Unknown
## 2 0 195.00 Used 0 None Unknown
## 3 0 219.99 Used 0 None Unknown
## 4 1 100.00 Used 0 None Unknown
## 5 0 210.99 Manufacturer refurbished 0 None Black
## 6 0 514.95 New other (see details) 0 None Gold
## storage productline UniqueID
## 1 32 iPad 1 11862
## 2 16 iPad mini 2 11863
## 3 64 iPad 3 11864
## 4 16 iPad mini 11865
## 5 32 iPad 3 11866
## 6 64 iPad Air 2 11867
## description
## 1 like new
## 142 iPad mini 1st gen wi-fi 16gb is in perfect working order.
## 309 In excellent condition. Minor scratches on the back. Screen in mint condition. Comes in original
## 312 iPad is in Great condition, the screen is in great condition showing only a few minor scratches, the
## 320 Good condition and fully functional
## 369
## biddable startprice condition cellular carrier color storage
## 1 0 105.00 Used 1 AT&T Unknown 32
## 142 1 0.99 Used 0 None Unknown 16
## 309 0 200.00 Used 1 AT&T Black 32
## 312 1 0.99 Used 0 None Unknown 16
## 320 1 60.00 Used 0 None White 16
## 369 1 197.97 Used 0 None Unknown 64
## productline UniqueID
## 1 iPad 1 11862
## 142 iPad mini 12003
## 309 iPad 3 12170
## 312 iPad mini 2 12173
## 320 iPad 1 12181
## 369 iPad mini 3 12230
## description
## 793 Crack on digitizer near top. Top line of digitizer does not respond to touch. Other than that, all
## 794
## 795
## 796
## 797
## 798 Slightly Used. Includes everything you need plus a nice leather case!\nThere is a slice mark on the
## biddable startprice condition cellular carrier color
## 793 0 104.00 For parts or not working 1 Unknown Black
## 794 0 95.00 Used 1 AT&T Unknown
## 795 1 199.99 Manufacturer refurbished 0 None White
## 796 0 149.99 Used 0 None Unknown
## 797 0 7.99 New Unknown Unknown Unknown
## 798 0 139.00 Used 1 Unknown Black
## storage productline UniqueID
## 793 16 iPad 2 12654
## 794 64 iPad 1 12655
## 795 16 iPad 4 12656
## 796 16 iPad 2 12657
## 797 Unknown iPad 3 12658
## 798 32 Unknown 12659
## 'data.frame': 798 obs. of 10 variables:
## $ description: chr "like new" "Item is in great shape. I upgraded to the iPad Air 2 and don't need the mini any longer, even though " "This iPad is working and is tested 100%. It runs great. It is in good condition. Cracked digitizer." "" ...
## $ biddable : int 0 0 0 1 0 0 0 0 0 1 ...
## $ startprice : num 105 195 220 100 211 ...
## $ condition : chr "Used" "Used" "Used" "Used" ...
## $ cellular : chr "1" "0" "0" "0" ...
## $ carrier : chr "AT&T" "None" "None" "None" ...
## $ color : chr "Unknown" "Unknown" "Unknown" "Unknown" ...
## $ storage : chr "32" "16" "64" "16" ...
## $ productline: chr "iPad 1" "iPad mini 2" "iPad 3" "iPad mini" ...
## $ UniqueID : int 11862 11863 11864 11865 11866 11867 11868 11869 11870 11871 ...
## - attr(*, "comment")= chr "glb_newobs_df"
## NULLif ((num_nas <- sum(is.na(glb_trnobs_df[, glb_rsp_var_raw]))) > 0)
stop("glb_trnobs_df$", glb_rsp_var_raw, " contains NAs for ", num_nas, " obs")
if (nrow(glb_trnobs_df) == nrow(glb_allobs_df))
warning("glb_trnobs_df same as glb_allobs_df")
if (nrow(glb_newobs_df) == nrow(glb_allobs_df))
warning("glb_newobs_df same as glb_allobs_df")
if (length(glb_drop_vars) > 0) {
warning("dropping vars: ", paste0(glb_drop_vars, collapse=", "))
glb_allobs_df <- glb_allobs_df[, setdiff(names(glb_allobs_df), glb_drop_vars)]
glb_trnobs_df <- glb_trnobs_df[, setdiff(names(glb_trnobs_df), glb_drop_vars)]
glb_newobs_df <- glb_newobs_df[, setdiff(names(glb_newobs_df), glb_drop_vars)]
}
#stop(here"); sav_allobs_df <- glb_allobs_df # glb_allobs_df <- sav_allobs_df
# Combine trnent & newobs into glb_allobs_df for easier manipulation
glb_trnobs_df$.src <- "Train"; glb_newobs_df$.src <- "Test";
glbFeatsExclude <- union(glbFeatsExclude, ".src")
glb_allobs_df <- myrbind_df(glb_trnobs_df, glb_newobs_df)
comment(glb_allobs_df) <- "glb_allobs_df"
# Check for duplicates in glb_id_var
if (length(glb_id_var) == 0) {
warning("using .rownames as identifiers for observations")
glb_allobs_df$.rownames <- rownames(glb_allobs_df)
glb_trnobs_df$.rownames <- rownames(subset(glb_allobs_df, .src == "Train"))
glb_newobs_df$.rownames <- rownames(subset(glb_allobs_df, .src == "Test"))
glb_id_var <- ".rownames"
}
if (sum(duplicated(glb_allobs_df[, glb_id_var, FALSE])) > 0)
stop(glb_id_var, " duplicated in glb_allobs_df")
glbFeatsExclude <- union(glbFeatsExclude, glb_id_var)
glb_allobs_df <- orderBy(reformulate(glb_id_var), glb_allobs_df)
glb_trnobs_df <- glb_newobs_df <- NULL
# For Tableau
write.csv(glb_allobs_df, "data/eBayiPadAll.csv", row.names=FALSE)
#stop(here"); glb_to_sav()
# Make any data corrections here
glb_allobs_df[glb_allobs_df[, glb_id_var] == 10986, "cellular"] <- "1"
glb_allobs_df[glb_allobs_df[, glb_id_var] == 10986, "carrier"] <- "T-Mobile"
# - Merge glb_obs_stack_condition & glbObsDropCondition
# - Derive glb_obs_stack|drop_chk_vars from condition automatically
# - Implement glb_obs_stack_condition & glb_obs_stack_chk_vars options
dsp_partition_stats <- function(obs_df, vars=NULL) {
lcl_vars <- NULL
for (var in c(vars, glb_rsp_var_raw)) {
if ((length(unique(obs_df[, var])) > 5) && is.numeric(obs_df[, var])) {
cut_var <- paste0(var, ".cut.fctr")
obs_df[, cut_var] <- cut(obs_df[, var], 3)
lcl_vars <- union(lcl_vars, cut_var)
} else lcl_vars <- union(lcl_vars, var)
}
print("Partition stats:")
print(mycreate_sqlxtab_df(obs_df, union(lcl_vars, ".src")))
for (var in lcl_vars) {
print(freq_df <- mycreate_sqlxtab_df(obs_df, union(var, ".src")))
print(myplot_hbar(freq_df, ".src", ".n", colorcol_name=var))
}
print(mycreate_sqlxtab_df(obs_df, ".src"))
# if (length(unique(glb_allobs_df[, glb_rsp_var_raw])) > 5) {
# cut_var <- paste0(glb_rsp_var_raw, ".cut.fctr")
# glb_allobs_df[, cut_var] <- cut(glb_allobs_df[, glb_rsp_var_raw], 3)
# glbFeatsExclude <- union(glbFeatsExclude, cut_var)
# glb_obs_stack_chk_vars <- union(cut_var, glb_obs_stack_chk_vars)
# } else glb_obs_stack_chk_vars <- union(glb_rsp_var_raw, glb_obs_stack_chk_vars)
# #glb_obs_stack_chk_vars <- union(glb_obs_stack_chk_vars, ".src")
# print(mycreate_sqlxtab_df(glb_allobs_df, union(var, ".src")))
# print(mycreate_sqlxtab_df(glb_allobs_df, union(glb_obs_stack_chk_vars, ".src")))
# for (var in glb_obs_stack_chk_vars) {
# print(mycreate_sqlxtab_df(glb_allobs_df, union(var, ".src")))
# }
# print(mycreate_sqlxtab_df(glb_allobs_df, ".src"))
}
myget_symbols <- function(txt) {
if (is.null(txt)) return(NULL)
#print(getParseData(parse(text=txt, keep.source=TRUE)))
return(unique(subset(getParseData(parse(text=txt, keep.source=TRUE)),
token == "SYMBOL")$text))
}
# tokens <- unlist(strsplit(gsub("[[:punct:]|[:space:]]", " ", glbObsDropCondition), " "))
# tokens <- tokens[tokens != ""]
# glb_obs_drop_chk_vars <- c("biddable") # or NULL
dsp_partition_stats(obs_df=glb_allobs_df, vars=myget_symbols(glbObsDropCondition))## [1] "Partition stats:"## Loading required package: sqldf
## Loading required package: gsubfn
## Loading required package: proto
## Loading required package: RSQLite
## Loading required package: DBI
## Loading required package: tcltk## UniqueID.cut.fctr productline sold .src .n
## 1 (1.18e+04,1.27e+04] iPad 2 NA Test 154
## 2 (1.18e+04,1.27e+04] iPad mini NA Test 111
## 3 (1e+04,1.09e+04] iPad 2 1 Train 104
## 4 (1.18e+04,1.27e+04] Unknown NA Test 92
## 5 (1.09e+04,1.18e+04] iPad mini 0 Train 91
## 6 (1e+04,1.09e+04] iPad 1 1 Train 91
## 7 (1.18e+04,1.27e+04] iPad 1 NA Test 88
## 8 (1.09e+04,1.18e+04] Unknown 0 Train 84
## 9 (1.09e+04,1.18e+04] iPad 2 0 Train 82
## 10 (1.18e+04,1.27e+04] iPad Air NA Test 74
## 11 (1e+04,1.09e+04] iPad mini 1 Train 74
## 12 (1.18e+04,1.27e+04] iPad 4 NA Test 68
## 13 (1.09e+04,1.18e+04] iPad Air 2 0 Train 65
## 14 (1.18e+04,1.27e+04] iPad Air 2 NA Test 62
## 15 (1e+04,1.09e+04] iPad 3 1 Train 61
## 16 (1.09e+04,1.18e+04] iPad Air 0 Train 58
## 17 (1.18e+04,1.27e+04] iPad mini 2 NA Test 56
## 18 (1.18e+04,1.27e+04] iPad 3 NA Test 55
## 19 (1.09e+04,1.18e+04] iPad mini 1 Train 54
## 20 (1e+04,1.09e+04] iPad 2 0 Train 52
## 21 (1e+04,1.09e+04] iPad Air 1 Train 51
## 22 (1.09e+04,1.18e+04] iPad 1 0 Train 50
## 23 (1e+04,1.09e+04] iPad mini 0 Train 48
## 24 (1.09e+04,1.18e+04] Unknown 1 Train 47
## 25 (1.09e+04,1.18e+04] iPad 4 0 Train 45
## 26 (1e+04,1.09e+04] iPad 1 0 Train 44
## 27 (1e+04,1.09e+04] iPad 4 0 Train 43
## 28 (1.09e+04,1.18e+04] iPad mini 3 0 Train 41
## 29 (1.09e+04,1.18e+04] iPad 2 1 Train 40
## 30 (1e+04,1.09e+04] iPad 4 1 Train 40
## 31 (1e+04,1.09e+04] iPad Air 0 Train 39
## 32 (1.18e+04,1.27e+04] iPad mini 3 NA Test 38
## 33 (1.09e+04,1.18e+04] iPad mini 2 0 Train 36
## 34 (1e+04,1.09e+04] iPad Air 2 1 Train 36
## 35 (1e+04,1.09e+04] iPad 3 0 Train 35
## 36 (1.09e+04,1.18e+04] iPad 3 0 Train 34
## 37 (1.09e+04,1.18e+04] iPad Air 2 1 Train 33
## 38 (1e+04,1.09e+04] Unknown 0 Train 29
## 39 (1e+04,1.09e+04] Unknown 1 Train 29
## 40 (1.09e+04,1.18e+04] iPad 1 1 Train 28
## 41 (1e+04,1.09e+04] iPad Air 2 0 Train 28
## 42 (1e+04,1.09e+04] iPad mini 2 1 Train 28
## 43 (1.09e+04,1.18e+04] iPad Air 1 Train 25
## 44 (1.09e+04,1.18e+04] iPad mini 2 1 Train 20
## 45 (1e+04,1.09e+04] iPad mini 2 0 Train 20
## 46 (1.09e+04,1.18e+04] iPad 4 1 Train 19
## 47 (1.09e+04,1.18e+04] iPad 3 1 Train 17
## 48 (1e+04,1.09e+04] iPad mini 3 0 Train 17
## 49 (1e+04,1.09e+04] iPad mini 3 1 Train 15
## 50 (1.09e+04,1.18e+04] iPad mini 3 1 Train 12
## 51 (1.18e+04,1.27e+04] Unknown 0 Train 9
## 52 (1.18e+04,1.27e+04] iPad 1 0 Train 8
## 53 (1.18e+04,1.27e+04] iPad Air 2 0 Train 7
## 54 (1.18e+04,1.27e+04] Unknown 1 Train 6
## 55 (1.18e+04,1.27e+04] iPad 1 1 Train 6
## 56 (1.18e+04,1.27e+04] iPad mini 0 Train 6
## 57 (1.18e+04,1.27e+04] iPad 2 0 Train 5
## 58 (1.18e+04,1.27e+04] iPad 4 0 Train 5
## 59 (1.18e+04,1.27e+04] iPad 4 1 Train 5
## 60 (1.18e+04,1.27e+04] iPad Air 0 Train 5
## 61 (1.18e+04,1.27e+04] iPad mini 3 0 Train 5
## 62 (1.18e+04,1.27e+04] iPad 3 0 Train 4
## 63 (1.18e+04,1.27e+04] iPad mini 1 Train 4
## 64 (1.09e+04,1.18e+04] iPad mini Retina 0 Train 3
## 65 (1.18e+04,1.27e+04] iPad 2 1 Train 3
## 66 (1.18e+04,1.27e+04] iPad 3 1 Train 2
## 67 (1.18e+04,1.27e+04] iPad Air 1 Train 2
## 68 (1.18e+04,1.27e+04] iPad Air 2 1 Train 2
## 69 (1.18e+04,1.27e+04] iPad mini 2 0 Train 2
## 70 (1e+04,1.09e+04] iPad mini Retina 1 Train 2
## 71 (1.09e+04,1.18e+04] iPad 5 1 Train 1
## 72 (1.09e+04,1.18e+04] iPad mini Retina 1 Train 1
## 73 (1.18e+04,1.27e+04] iPad mini 2 1 Train 1
## 74 (1.18e+04,1.27e+04] iPad mini Retina 1 Train 1
## 75 (1e+04,1.09e+04] iPad mini Retina 0 Train 1
## UniqueID.cut.fctr .src .n
## 1 (1e+04,1.09e+04] Train 887
## 2 (1.09e+04,1.18e+04] Train 886
## 3 (1.18e+04,1.27e+04] Test 798
## 4 (1.18e+04,1.27e+04] Train 88
## productline .src .n
## 1 iPad 2 Train 286
## 2 iPad mini Train 277
## 3 iPad 1 Train 227
## 4 Unknown Train 204
## 5 iPad Air Train 180
## 6 iPad Air 2 Train 171
## 7 iPad 4 Train 157
## 8 iPad 2 Test 154
## 9 iPad 3 Train 153
## 10 iPad mini Test 111
## 11 iPad mini 2 Train 107
## 12 Unknown Test 92
## 13 iPad mini 3 Train 90
## 14 iPad 1 Test 88
## 15 iPad Air Test 74
## 16 iPad 4 Test 68
## 17 iPad Air 2 Test 62
## 18 iPad mini 2 Test 56
## 19 iPad 3 Test 55
## 20 iPad mini 3 Test 38
## 21 iPad mini Retina Train 8
## 22 iPad 5 Train 1
## sold .src .n
## 1 0 Train 1001
## 2 1 Train 860
## 3 NA Test 798
## .src .n
## 1 Train 1861
## 2 Test 798if (!is.null(glbObsDropCondition)) {
print(sprintf("Running glbObsDropCondition filter: %s", glbObsDropCondition))
glb_allobs_df <- do.call("subset",
list(glb_allobs_df, parse(text=paste0("!(", glbObsDropCondition, ")"))))
dsp_partition_stats(obs_df=glb_allobs_df, vars=myget_symbols(glbObsDropCondition))
}## [1] "Running glbObsDropCondition filter: (UniqueID %in% c(NULL\n , 11234 #sold=0; 2 other dups(10306, 11503) are sold=1\n , 11844 #sold=0; 3 other dups(11721, 11738, 11812) are sold=1\n )) | \n (productline %in% c('iPad 5', 'iPad mini Retina'))\n # | (biddable != 0) # bid0_sp\n # | (biddable == 0) # bid1_sp\n "
## [1] "Partition stats:"
## UniqueID.cut.fctr productline sold .src .n
## 1 (1.18e+04,1.27e+04] iPad 2 NA Test 154
## 2 (1.18e+04,1.27e+04] iPad mini NA Test 111
## 3 (1e+04,1.09e+04] iPad 2 1 Train 104
## 4 (1.18e+04,1.27e+04] Unknown NA Test 92
## 5 (1.09e+04,1.18e+04] iPad mini 0 Train 91
## 6 (1e+04,1.09e+04] iPad 1 1 Train 91
## 7 (1.18e+04,1.27e+04] iPad 1 NA Test 88
## 8 (1.09e+04,1.18e+04] Unknown 0 Train 84
## 9 (1.09e+04,1.18e+04] iPad 2 0 Train 82
## 10 (1.18e+04,1.27e+04] iPad Air NA Test 74
## 11 (1e+04,1.09e+04] iPad mini 1 Train 74
## 12 (1.18e+04,1.27e+04] iPad 4 NA Test 68
## 13 (1.09e+04,1.18e+04] iPad Air 2 0 Train 65
## 14 (1.18e+04,1.27e+04] iPad Air 2 NA Test 62
## 15 (1e+04,1.09e+04] iPad 3 1 Train 61
## 16 (1.09e+04,1.18e+04] iPad Air 0 Train 58
## 17 (1.18e+04,1.27e+04] iPad mini 2 NA Test 56
## 18 (1.18e+04,1.27e+04] iPad 3 NA Test 55
## 19 (1.09e+04,1.18e+04] iPad mini 1 Train 54
## 20 (1e+04,1.09e+04] iPad 2 0 Train 52
## 21 (1e+04,1.09e+04] iPad Air 1 Train 51
## 22 (1.09e+04,1.18e+04] iPad 1 0 Train 49
## 23 (1e+04,1.09e+04] iPad mini 0 Train 48
## 24 (1.09e+04,1.18e+04] Unknown 1 Train 47
## 25 (1.09e+04,1.18e+04] iPad 4 0 Train 45
## 26 (1e+04,1.09e+04] iPad 1 0 Train 44
## 27 (1e+04,1.09e+04] iPad 4 0 Train 43
## 28 (1.09e+04,1.18e+04] iPad mini 3 0 Train 41
## 29 (1.09e+04,1.18e+04] iPad 2 1 Train 40
## 30 (1e+04,1.09e+04] iPad 4 1 Train 40
## 31 (1e+04,1.09e+04] iPad Air 0 Train 39
## 32 (1.18e+04,1.27e+04] iPad mini 3 NA Test 38
## 33 (1.09e+04,1.18e+04] iPad mini 2 0 Train 36
## 34 (1e+04,1.09e+04] iPad Air 2 1 Train 36
## 35 (1e+04,1.09e+04] iPad 3 0 Train 35
## 36 (1.09e+04,1.18e+04] iPad 3 0 Train 34
## 37 (1.09e+04,1.18e+04] iPad Air 2 1 Train 33
## 38 (1e+04,1.09e+04] Unknown 0 Train 29
## 39 (1e+04,1.09e+04] Unknown 1 Train 29
## 40 (1.09e+04,1.18e+04] iPad 1 1 Train 28
## 41 (1e+04,1.09e+04] iPad Air 2 0 Train 28
## 42 (1e+04,1.09e+04] iPad mini 2 1 Train 28
## 43 (1.09e+04,1.18e+04] iPad Air 1 Train 25
## 44 (1.09e+04,1.18e+04] iPad mini 2 1 Train 20
## 45 (1e+04,1.09e+04] iPad mini 2 0 Train 20
## 46 (1.09e+04,1.18e+04] iPad 4 1 Train 19
## 47 (1.09e+04,1.18e+04] iPad 3 1 Train 17
## 48 (1e+04,1.09e+04] iPad mini 3 0 Train 17
## 49 (1e+04,1.09e+04] iPad mini 3 1 Train 15
## 50 (1.09e+04,1.18e+04] iPad mini 3 1 Train 12
## 51 (1.18e+04,1.27e+04] Unknown 0 Train 9
## 52 (1.18e+04,1.27e+04] iPad 1 0 Train 7
## 53 (1.18e+04,1.27e+04] iPad Air 2 0 Train 7
## 54 (1.18e+04,1.27e+04] Unknown 1 Train 6
## 55 (1.18e+04,1.27e+04] iPad 1 1 Train 6
## 56 (1.18e+04,1.27e+04] iPad mini 0 Train 6
## 57 (1.18e+04,1.27e+04] iPad 2 0 Train 5
## 58 (1.18e+04,1.27e+04] iPad 4 0 Train 5
## 59 (1.18e+04,1.27e+04] iPad 4 1 Train 5
## 60 (1.18e+04,1.27e+04] iPad Air 0 Train 5
## 61 (1.18e+04,1.27e+04] iPad mini 3 0 Train 5
## 62 (1.18e+04,1.27e+04] iPad 3 0 Train 4
## 63 (1.18e+04,1.27e+04] iPad mini 1 Train 4
## 64 (1.18e+04,1.27e+04] iPad 2 1 Train 3
## 65 (1.18e+04,1.27e+04] iPad 3 1 Train 2
## 66 (1.18e+04,1.27e+04] iPad Air 1 Train 2
## 67 (1.18e+04,1.27e+04] iPad Air 2 1 Train 2
## 68 (1.18e+04,1.27e+04] iPad mini 2 0 Train 2
## 69 (1.18e+04,1.27e+04] iPad mini 2 1 Train 1
## UniqueID.cut.fctr .src .n
## 1 (1e+04,1.09e+04] Train 884
## 2 (1.09e+04,1.18e+04] Train 880
## 3 (1.18e+04,1.27e+04] Test 798
## 4 (1.18e+04,1.27e+04] Train 86
## productline .src .n
## 1 iPad 2 Train 286
## 2 iPad mini Train 277
## 3 iPad 1 Train 225
## 4 Unknown Train 204
## 5 iPad Air Train 180
## 6 iPad Air 2 Train 171
## 7 iPad 4 Train 157
## 8 iPad 2 Test 154
## 9 iPad 3 Train 153
## 10 iPad mini Test 111
## 11 iPad mini 2 Train 107
## 12 Unknown Test 92
## 13 iPad mini 3 Train 90
## 14 iPad 1 Test 88
## 15 iPad Air Test 74
## 16 iPad 4 Test 68
## 17 iPad Air 2 Test 62
## 18 iPad mini 2 Test 56
## 19 iPad 3 Test 55
## 20 iPad mini 3 Test 38
## sold .src .n
## 1 0 Train 995
## 2 1 Train 855
## 3 NA Test 798
## .src .n
## 1 Train 1850
## 2 Test 798# Check for duplicates by all features
require(gdata)## Loading required package: gdata
## gdata: read.xls support for 'XLS' (Excel 97-2004) files ENABLED.
##
## gdata: read.xls support for 'XLSX' (Excel 2007+) files ENABLED.
##
## Attaching package: 'gdata'
##
## The following object is masked from 'package:stats':
##
## nobs
##
## The following object is masked from 'package:utils':
##
## object.size#print(names(glb_allobs_df))
dup_allobs_df <- glb_allobs_df[duplicated2(subset(glb_allobs_df,
select=-c(UniqueID, sold, .src))), ]
dup_allobs_df <- orderBy(~productline+description+startprice+biddable, dup_allobs_df)
print(sprintf("Found %d duplicates by all features:", nrow(dup_allobs_df)))## [1] "Found 304 duplicates by all features:"myprint_df(dup_allobs_df)## description biddable startprice condition cellular
## 1711 1 0.99 For parts or not working Unknown
## 2608 1 0.99 For parts or not working Unknown
## 293 1 5.00 Used Unknown
## 478 1 5.00 Used Unknown
## 385 0 15.00 Used 0
## 390 0 15.00 Used 0
## carrier color storage productline sold UniqueID .src
## 1711 Unknown Unknown 16 Unknown 1 11711 Train
## 2608 Unknown Unknown 16 Unknown NA 12608 Test
## 293 Unknown White 16 Unknown 1 10293 Train
## 478 Unknown White 16 Unknown 1 10478 Train
## 385 None Black 16 Unknown 0 10385 Train
## 390 None Black 16 Unknown 0 10390 Train
## description biddable startprice condition cellular
## 1956 1 0.99 Used 0
## 828 1 249.97 Manufacturer refurbished 1
## 3 0 199.99 Used 0
## 1649 0 209.00 For parts or not working Unknown
## 2111 1 200.00 Used 0
## 172 0 269.00 Used 0
## carrier color storage productline sold UniqueID .src
## 1956 None Unknown 16 iPad 2 NA 11956 Test
## 828 Unknown Black 64 iPad 2 0 10828 Train
## 3 None White 16 iPad 4 1 10003 Train
## 1649 Unknown Unknown 16 iPad Air 0 11649 Train
## 2111 None Space Gray 64 iPad mini 2 NA 12111 Test
## 172 None Unknown 32 iPad mini 2 0 10172 Train
## description biddable startprice condition cellular carrier color
## 8 0 329.99 New 0 None White
## 660 0 329.99 New 0 None White
## 319 0 345.00 New 0 None Gold
## 1886 0 345.00 New 0 None Gold
## 1363 0 498.88 New 1 Verizon Gold
## 1394 0 498.88 New 1 Verizon Gold
## storage productline sold UniqueID .src
## 8 16 iPad mini 3 0 10008 Train
## 660 16 iPad mini 3 0 10660 Train
## 319 16 iPad mini 3 1 10319 Train
## 1886 16 iPad mini 3 NA 11886 Test
## 1363 16 iPad mini 3 0 11363 Train
## 1394 16 iPad mini 3 0 11394 Train# print(dup_allobs_df[, c(glb_id_var, glb_rsp_var_raw,
# "description", "startprice", "biddable")])
# write.csv(dup_allobs_df[, c("UniqueID"), FALSE], "ebayipads_dups.csv", row.names=FALSE)
dupobs_df <- tidyr::unite(dup_allobs_df, "allfeats", -c(sold, UniqueID, .src), sep="#")
# dupobs_df <- dplyr::group_by(dupobs_df, allfeats)
# dupobs_df <- dupobs_df[, "UniqueID", FALSE]
# dupobs_df <- ungroup(dupobs_df)
#
# dupobs_df$.rownames <- row.names(dupobs_df)
grpobs_df <- data.frame(allfeats=unique(dupobs_df[, "allfeats"]))
grpobs_df$.grpid <- row.names(grpobs_df)
dupobs_df <- merge(dupobs_df, grpobs_df)
# dupobs_tbl <- table(dupobs_df$.grpid)
# print(max(dupobs_tbl))
# print(dupobs_tbl[which.max(dupobs_tbl)])
# print(dupobs_df[dupobs_df$.grpid == names(dupobs_tbl[which.max(dupobs_tbl)]), ])
# print(dupobs_df[dupobs_df$.grpid == 106, ])
# for (grpid in c(9, 17, 31, 36, 53))
# print(dupobs_df[dupobs_df$.grpid == grpid, ])
dupgrps_df <- as.data.frame(table(dupobs_df$.grpid, dupobs_df$sold, useNA="ifany"))
names(dupgrps_df)[c(1,2)] <- c(".grpid", "sold")
dupgrps_df$.grpid <- as.numeric(as.character(dupgrps_df$.grpid))
dupgrps_df <- tidyr::spread(dupgrps_df, sold, Freq)
names(dupgrps_df)[-1] <- paste("sold", names(dupgrps_df)[-1], sep=".")
dupgrps_df$.freq <- sapply(1:nrow(dupgrps_df), function(row) sum(dupgrps_df[row, -1]))
myprint_df(orderBy(~-.freq, dupgrps_df))## .grpid sold.0 sold.1 sold.NA .freq
## 40 40 0 6 3 9
## 106 106 0 4 1 5
## 9 9 0 1 3 4
## 17 17 0 3 1 4
## 36 36 0 3 1 4
## 53 53 0 2 2 4
## .grpid sold.0 sold.1 sold.NA .freq
## 10 10 0 2 0 2
## 42 42 0 1 1 2
## 57 57 1 0 1 2
## 66 66 1 0 1 2
## 91 91 0 1 1 2
## 101 101 0 1 1 2
## .grpid sold.0 sold.1 sold.NA .freq
## 130 130 1 0 1 2
## 131 131 1 1 0 2
## 132 132 0 1 1 2
## 133 133 2 0 0 2
## 134 134 0 1 1 2
## 135 135 2 0 0 2print("sold Conflicts:")## [1] "sold Conflicts:"print(subset(dupgrps_df, (sold.0 > 0) & (sold.1 > 0)))## .grpid sold.0 sold.1 sold.NA .freq
## 4 4 1 1 0 2
## 22 22 1 1 0 2
## 23 23 1 1 0 2
## 74 74 1 1 0 2
## 83 83 1 1 0 2
## 84 84 1 1 0 2
## 95 95 1 1 0 2
## 102 102 1 1 0 2
## 109 109 1 1 0 2
## 111 111 1 1 0 2
## 122 122 1 1 0 2
## 131 131 1 1 0 2#dupobs_df[dupobs_df$.grpid == 4, ]
glb_allobs_df <- merge(glb_allobs_df, dupobs_df[, c(glb_id_var, ".grpid")],
by=glb_id_var, all.x=TRUE)
if (nrow(subset(dupgrps_df, (sold.0 > 0) & (sold.1 > 0) & (sold.0 != sold.1))) > 0)
stop("Duplicate conflicts are resolvable")
#subset(glb_allobs_df, .grpid %in% c(25))
#mydsp_obs(list(productline.contains="iPad 1", storage.contains="16", color.contains="Black", carrier.contains="None", cellular.contains="0", condition.contains="Used", startprice=80), cols=c("productline", "storage", "color", "carrier", "cellular", "condition", "startprice", "sold"))
print("Test & Train Groups:")## [1] "Test & Train Groups:"print(subset(dupgrps_df, (sold.NA > 0)))## .grpid sold.0 sold.1 sold.NA .freq
## 1 1 0 1 1 2
## 5 5 1 0 1 2
## 7 7 0 0 2 2
## 8 8 1 0 1 2
## 9 9 0 1 3 4
## 12 12 0 0 2 2
## 14 14 0 1 1 2
## 15 15 0 0 2 2
## 17 17 0 3 1 4
## 18 18 0 2 1 3
## 19 19 0 2 1 3
## 24 24 0 2 1 3
## 26 26 1 0 1 2
## 28 28 1 0 1 2
## 30 30 0 1 1 2
## 32 32 0 0 2 2
## 33 33 0 1 1 2
## 35 35 0 2 1 3
## 36 36 0 3 1 4
## 37 37 0 0 2 2
## 38 38 0 1 1 2
## 40 40 0 6 3 9
## 41 41 0 0 2 2
## 42 42 0 1 1 2
## 43 43 0 1 1 2
## 44 44 0 2 1 3
## 47 47 0 1 1 2
## 48 48 0 0 2 2
## 49 49 0 1 2 3
## 51 51 0 1 1 2
## 53 53 0 2 2 4
## 54 54 0 1 1 2
## 55 55 1 0 2 3
## 56 56 1 0 1 2
## 57 57 1 0 1 2
## 58 58 0 0 2 2
## 59 59 1 0 1 2
## 60 60 1 0 1 2
## 63 63 0 1 1 2
## 66 66 1 0 1 2
## 67 67 1 0 1 2
## 68 68 0 0 2 2
## 69 69 1 0 1 2
## 73 73 0 1 1 2
## 76 76 0 2 1 3
## 86 86 0 0 2 2
## 87 87 1 0 1 2
## 89 89 1 0 1 2
## 90 90 0 0 2 2
## 91 91 0 1 1 2
## 93 93 0 1 1 2
## 94 94 1 0 1 2
## 99 99 0 1 1 2
## 101 101 0 1 1 2
## 103 103 0 1 1 2
## 104 104 1 0 1 2
## 106 106 0 4 1 5
## 107 107 0 1 1 2
## 108 108 0 1 1 2
## 112 112 1 0 1 2
## 114 114 0 1 1 2
## 115 115 0 1 1 2
## 116 116 1 0 1 2
## 117 117 0 2 1 3
## 118 118 0 1 1 2
## 121 121 1 0 1 2
## 124 124 1 0 1 2
## 128 128 0 1 1 2
## 130 130 1 0 1 2
## 132 132 0 1 1 2
## 134 134 0 1 1 2glbFeatsExclude <- c(".grpid", glbFeatsExclude)
#stop(here"); glb_to_sav(); all.equal(sav_allobs_df, glb_allobs_df); glb_allobs_df <- sav_allobs_df
if (!is.null(glbInpMerge)) {
print("Running glbInpMerge specs...")
obsMrg <- data.frame()
for (fName in glbInpMerge$fnames) {
print(sprintf(" Appending rows from %s...", fName))
obsMrg <- rbind(obsMrg, read.csv(fName))
}
glb_allobs_df <- merge(glb_allobs_df, obsMrg, all.x = TRUE)
}## [1] "Running glbInpMerge specs..."
## [1] " Appending rows from ebayipads_finmdl_bid0_sp_out.csv..."
## [1] " Appending rows from ebayipads_mdlens_bid1_sp_out.csv..."dsp_partition_stats(obs_df = glb_allobs_df,
vars = myget_symbols(glb_obs_repartition_train_condition))## [1] "Partition stats:"
## sold .src .n
## 1 0 Train 995
## 2 1 Train 855
## 3 NA Test 798
## sold .src .n
## 1 0 Train 995
## 2 1 Train 855
## 3 NA Test 798
## .src .n
## 1 Train 1850
## 2 Test 798if (!is.null(glb_obs_repartition_train_condition)) {
print(sprintf("Running glb_obs_repartition_train_condition filter: %s",
glb_obs_repartition_train_condition))
# glb_allobs_df <- mutate(glb_allobs_df, .src=ifelse(!is.na(sold) & (sold == 1),
# "Train", "Test"))
# glb_allobs_df <- mutate_(glb_allobs_df,
# .src=interp(ifelse(eval(parse(text="!is.na(sold) & (sold == 1)")),
# "Train", "Test")))
# glb_allobs_df <- within(glb_allobs_df, {
# .src <- ifelse(eval(parse(text="!is.na(sold) & (sold == 1)")),
# "Train", "Test")
# })
# glb_allobs_df <- within(glb_allobs_df, {
# if(eval(parse(text="!is.na(sold) & (sold == 1)"))) .src <- "Train" else
# .src <- "Test"
# })
# with(glb_allobs_df, {
# src <- ifelse(eval(parse(text="!is.na(sold) & (sold == 1)")),
# "Train", "Test")
# })
# glb_allobs_df$.src <- sapply(1:nrow(glb_allobs_df), function (row_ix) ifelse)
# glb_allobs_df[parse(text=paste0("!(", glbObsDropCondition, ")")), ".src"] <- do.call("subset",
# list(glb_allobs_df, ))
glb_trnobs_df <- do.call("subset", list(glb_allobs_df,
parse(text=paste0(" (", glb_obs_repartition_train_condition, ")"))))
glb_trnobs_df$.src <- "Train"
glb_newobs_df <- do.call("subset", list(glb_allobs_df,
parse(text=paste0("!(", glb_obs_repartition_train_condition, ")"))))
glb_newobs_df$.src <- "Test"
glb_allobs_df <- rbind(glb_trnobs_df, glb_newobs_df)
dsp_partition_stats(obs_df = glb_allobs_df,
vars = myget_symbols(glb_obs_repartition_train_condition))
}
glb_chunks_df <- myadd_chunk(glb_chunks_df, "inspect.data", major.inc=TRUE)## label step_major step_minor label_minor bgn end elapsed
## 1 import.data 1 0 0 23.954 40.783 16.829
## 2 inspect.data 2 0 0 40.783 NA NAStep 2.0: inspect data
#print(str(glb_allobs_df))
#View(glb_allobs_df)
dsp_class_dstrb <- function(var) {
xtab_df <- mycreate_xtab_df(glb_allobs_df, c(".src", var))
rownames(xtab_df) <- xtab_df$.src
xtab_df <- subset(xtab_df, select=-.src)
print(xtab_df)
print(xtab_df / rowSums(xtab_df, na.rm=TRUE))
}
# Performed repeatedly in other chunks
glb_chk_data <- function(featsExclude = glbFeatsExclude,
fctrMaxUniqVals = glbFctrMaxUniqVals) {
# Histogram of predictor in glb_trnobs_df & glb_newobs_df
print(myplot_histogram(glb_allobs_df, glb_rsp_var_raw) + facet_wrap(~ .src))
if (glb_is_classification)
dsp_class_dstrb(var=ifelse(glb_rsp_var %in% names(glb_allobs_df),
glb_rsp_var, glb_rsp_var_raw))
mycheck_problem_data(glb_allobs_df, featsExclude, fctrMaxUniqVals)
}
glb_chk_data()## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.## Warning: Removed 798 rows containing non-finite values (stat_bin).## Loading required package: reshape2
## sold.0 sold.1 sold.NA
## Test NA NA 798
## Train 995 855 NA
## sold.0 sold.1 sold.NA
## Test NA NA 1
## Train 0.5378378 0.4621622 NA
## [1] "numeric data missing in : "
## sold
## 798
## [1] "numeric data w/ 0s in : "
## biddable sold
## 1437 995
## [1] "numeric data w/ Infs in : "
## named integer(0)
## [1] "numeric data w/ NaNs in : "
## named integer(0)
## [1] "string data missing in : "
## description condition cellular carrier color storage
## 1513 0 0 0 0 0
## productline .grpid
## 0 NA# Create new features that help diagnostics
if (!is.null(glb_map_rsp_raw_to_var)) {
glb_allobs_df[, glb_rsp_var] <-
glb_map_rsp_raw_to_var(glb_allobs_df[, glb_rsp_var_raw])
mycheck_map_results(mapd_df=glb_allobs_df,
from_col_name=glb_rsp_var_raw, to_col_name=glb_rsp_var)
if (glb_is_classification) dsp_class_dstrb(glb_rsp_var)
}## sold sold.fctr .n
## 1 0 N 995
## 2 1 Y 855
## 3 NA <NA> 798## Warning: Removed 1 rows containing missing values (position_stack).
## sold.fctr.N sold.fctr.Y sold.fctr.NA
## Test NA NA 798
## Train 995 855 NA
## sold.fctr.N sold.fctr.Y sold.fctr.NA
## Test NA NA 1
## Train 0.5378378 0.4621622 NA# check distribution of all numeric data
dsp_numeric_feats_dstrb <- function(feats_vctr) {
grid.newpage()
pushViewport(viewport(layout = grid.layout(ceiling(length(feats_vctr) / 2.0), 2)))
pltIx <- 1
for (feat in feats_vctr) {
#print(sprintf("feat: %s", feat))
if (glb_is_regression)
gp <- myplot_scatter(df=glb_allobs_df, ycol_name=glb_rsp_var, xcol_name=feat,
smooth=TRUE)
if (glb_is_classification)
#gp <- myplot_box(df=glb_allobs_df, ycol_names=feat, xcol_name=glb_rsp_var)
gp <- myplot_violin(glb_allobs_df, ycol_names = feat, xcol_name = glb_rsp_var)
if (inherits(glb_allobs_df[, feat], "factor"))
gp <- gp + facet_wrap(reformulate(feat))
print(gp + labs(title = feat),
vp = viewport(layout.pos.row = ceiling(pltIx / 2.0),
layout.pos.col = ((pltIx - 1) %% 2) + 1))
pltIx <- pltIx + 1
}
}
# dsp_numeric_feats_dstrb(setdiff(names(glb_allobs_df), union(myfind_chr_cols_df(glb_allobs_df), c(glb_rsp_var_raw, glb_rsp_var)))[2:6]) # dsp_numeric_feats_dstrb(c("startprice", "sprice.root3", "sprice.predict.diff"))
add_new_diag_feats <- function(obs_df, ref_df=glb_allobs_df) {
require(plyr)
set.seed(169)
obs_df <- mutate(obs_df,
# <col_name>.NA=is.na(<col_name>),
# <col_name>.fctr=factor(<col_name>,
# as.factor(union(obs_df$<col_name>, obs_twin_df$<col_name>))),
# <col_name>.fctr=relevel(factor(<col_name>,
# as.factor(union(obs_df$<col_name>, obs_twin_df$<col_name>))),
# "<ref_val>"),
# <col2_name>.fctr=relevel(factor(ifelse(<col1_name> == <val>, "<oth_val>", "<ref_val>")),
# as.factor(c("R", "<ref_val>")),
# ref="<ref_val>"),
# This doesn't work - use sapply instead
# <col_name>.fctr_num=grep(<col_name>, levels(<col_name>.fctr)),
#
# Date.my=as.Date(strptime(Date, "%m/%d/%y %H:%M")),
# Year=year(Date.my),
# Month=months(Date.my),
# Weekday=weekdays(Date.my)
# <col_name>=<table>[as.character(<col2_name>)],
# <col_name>=as.numeric(<col2_name>),
# <col_name> = trunc(<col2_name> / 100),
.rnorm = rnorm(n=nrow(obs_df))
)
# If levels of a factor are different across obs_df & glb_newobs_df; predict.glm fails
# Transformations not handled by mutate
# obs_df$<col_name>.fctr.num <- sapply(1:nrow(obs_df),
# function(row_ix) grep(obs_df[row_ix, "<col_name>"],
# levels(obs_df[row_ix, "<col_name>.fctr"])))
#print(summary(obs_df))
#print(sapply(names(obs_df), function(col) sum(is.na(obs_df[, col]))))
return(obs_df)
}
glb_allobs_df <- add_new_diag_feats(glb_allobs_df)## Loading required package: plyrrequire(dplyr)## Loading required package: dplyr
##
## Attaching package: 'dplyr'
##
## The following objects are masked from 'package:plyr':
##
## arrange, count, desc, failwith, id, mutate, rename, summarise,
## summarize
##
## The following objects are masked from 'package:gdata':
##
## combine, first, last
##
## The following objects are masked from 'package:stats':
##
## filter, lag
##
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union#stop(here"); sav_allobs_df <- glb_allobs_df # glb_allobs_df <- sav_allobs_df
# Merge some <descriptor>
# glb_allobs_df$<descriptor>.my <- glb_allobs_df$<descriptor>
# glb_allobs_df[grepl("\\bAIRPORT\\b", glb_allobs_df$<descriptor>.my),
# "<descriptor>.my"] <- "AIRPORT"
# Check distributions of newly transformed / extracted vars
# Enhancement: remove vars that were displayed ealier
dsp_numeric_feats_dstrb(feats_vctr=setdiff(names(glb_allobs_df),
c(myfind_chr_cols_df(glb_allobs_df), glb_rsp_var_raw, glb_rsp_var,
glbFeatsExclude)))
# Convert factors to dummy variables
# Build splines require(splines); bsBasis <- bs(training$age, df=3)
#pairs(subset(glb_trnobs_df, select=-c(col_symbol)))
# Check for glb_newobs_df & glb_trnobs_df features range mismatches
# Other diagnostics:
# print(subset(glb_trnobs_df, <col1_name> == max(glb_trnobs_df$<col1_name>, na.rm=TRUE) &
# <col2_name> <= mean(glb_trnobs_df$<col1_name>, na.rm=TRUE)))
# print(glb_trnobs_df[which.max(glb_trnobs_df$<col_name>),])
# print(<col_name>_freq_glb_trnobs_df <- mycreate_tbl_df(glb_trnobs_df, "<col_name>"))
# print(which.min(table(glb_trnobs_df$<col_name>)))
# print(which.max(table(glb_trnobs_df$<col_name>)))
# print(which.max(table(glb_trnobs_df$<col1_name>, glb_trnobs_df$<col2_name>)[, 2]))
# print(table(glb_trnobs_df$<col1_name>, glb_trnobs_df$<col2_name>))
# print(table(is.na(glb_trnobs_df$<col1_name>), glb_trnobs_df$<col2_name>))
# print(table(sign(glb_trnobs_df$<col1_name>), glb_trnobs_df$<col2_name>))
# print(mycreate_xtab_df(glb_trnobs_df, <col1_name>))
# print(mycreate_xtab_df(glb_trnobs_df, c(<col1_name>, <col2_name>)))
# print(<col1_name>_<col2_name>_xtab_glb_trnobs_df <-
# mycreate_xtab_df(glb_trnobs_df, c("<col1_name>", "<col2_name>")))
# <col1_name>_<col2_name>_xtab_glb_trnobs_df[is.na(<col1_name>_<col2_name>_xtab_glb_trnobs_df)] <- 0
# print(<col1_name>_<col2_name>_xtab_glb_trnobs_df <-
# mutate(<col1_name>_<col2_name>_xtab_glb_trnobs_df,
# <col3_name>=(<col1_name> * 1.0) / (<col1_name> + <col2_name>)))
# print(mycreate_sqlxtab_df(glb_allobs_df, c("<col1_name>", "<col2_name>")))
# print(<col2_name>_min_entity_arr <-
# sort(tapply(glb_trnobs_df$<col1_name>, glb_trnobs_df$<col2_name>, min, na.rm=TRUE)))
# print(<col1_name>_na_by_<col2_name>_arr <-
# sort(tapply(glb_trnobs_df$<col1_name>.NA, glb_trnobs_df$<col2_name>, mean, na.rm=TRUE)))
# Other plots:
# print(myplot_box(df=glb_trnobs_df, ycol_names="<col1_name>"))
# print(myplot_box(df=glb_trnobs_df, ycol_names="<col1_name>", xcol_name="<col2_name>"))
# print(myplot_line(subset(glb_trnobs_df, Symbol %in% c("CocaCola", "ProcterGamble")),
# "Date.POSIX", "StockPrice", facet_row_colnames="Symbol") +
# geom_vline(xintercept=as.numeric(as.POSIXlt("2003-03-01"))) +
# geom_vline(xintercept=as.numeric(as.POSIXlt("1983-01-01")))
# )
# print(myplot_line(subset(glb_trnobs_df, Date.POSIX > as.POSIXct("2004-01-01")),
# "Date.POSIX", "StockPrice") +
# geom_line(aes(color=Symbol)) +
# coord_cartesian(xlim=c(as.POSIXct("1990-01-01"),
# as.POSIXct("2000-01-01"))) +
# coord_cartesian(ylim=c(0, 250)) +
# geom_vline(xintercept=as.numeric(as.POSIXlt("1997-09-01"))) +
# geom_vline(xintercept=as.numeric(as.POSIXlt("1997-11-01")))
# )
# print(myplot_scatter(glb_allobs_df, "<col1_name>", "<col2_name>", smooth=TRUE))
# print(myplot_scatter(glb_allobs_df, "<col1_name>", "<col2_name>", colorcol_name="<Pred.fctr>") +
# geom_point(data=subset(glb_allobs_df, <condition>),
# mapping=aes(x=<x_var>, y=<y_var>), color="red", shape=4, size=5) +
# geom_vline(xintercept=84))
glb_chunks_df <- myadd_chunk(glb_chunks_df, "scrub.data", major.inc=FALSE)## label step_major step_minor label_minor bgn end elapsed
## 2 inspect.data 2 0 0 40.783 45.356 4.574
## 3 scrub.data 2 1 1 45.357 NA NAStep 2.1: scrub data
mycheck_problem_data(glb_allobs_df, featsExclude = glbFeatsExclude,
fctrMaxUniqVals = glbFctrMaxUniqVals)## [1] "numeric data missing in : "
## sold sold.fctr
## 798 798
## [1] "numeric data w/ 0s in : "
## biddable sold
## 1437 995
## [1] "numeric data w/ Infs in : "
## named integer(0)
## [1] "numeric data w/ NaNs in : "
## named integer(0)
## [1] "string data missing in : "
## description condition cellular carrier color storage
## 1513 0 0 0 0 0
## productline .grpid
## 0 NAfindOffendingCharacter <- function(x, maxStringLength=256){
print(x)
for (c in 1:maxStringLength){
offendingChar <- substr(x,c,c)
#print(offendingChar) #uncomment if you want the indiv characters printed
#the next character is the offending multibyte Character
}
}
# string_vector <- c("test", "Se\x96ora", "works fine")
# lapply(string_vector, findOffendingCharacter)
# lapply(glb_allobs_df$description[29], findOffendingCharacter)
dsp_hdlxtab <- function(str)
print(mycreate_sqlxtab_df(glb_allobs_df[sel_obs(Headline.contains=str), ],
c("Headline.pfx", "Headline", glb_rsp_var)))
#dsp_hdlxtab("(1914)|(1939)")
dsp_catxtab <- function(str)
print(mycreate_sqlxtab_df(glb_allobs_df[sel_obs(Headline.contains=str), ],
c("Headline.pfx", "NewsDesk", "SectionName", "SubsectionName", glb_rsp_var)))
# dsp_catxtab("1914)|(1939)")
# dsp_catxtab("19(14|39|64):")
# dsp_catxtab("19..:")
# Merge some categories
# glb_allobs_df$myCategory <-
# plyr::revalue(glb_allobs_df$myCategory, c(
# "#Business Day#Dealbook" = "Business#Business Day#Dealbook",
# "#Business Day#Small Business" = "Business#Business Day#Small Business",
# "dummy" = "dummy"
# ))
# ctgry_xtab_df <- orderBy(reformulate(c("-", ".n")),
# mycreate_sqlxtab_df(glb_allobs_df,
# c("myCategory", "NewsDesk", "SectionName", "SubsectionName", glb_rsp_var)))
# myprint_df(ctgry_xtab_df)
# write.table(ctgry_xtab_df, paste0(glb_out_pfx, "ctgry_xtab.csv"),
# row.names=FALSE)
# ctgry_cast_df <- orderBy(~ -Y -NA, dcast(ctgry_xtab_df,
# myCategory + NewsDesk + SectionName + SubsectionName ~
# Popular.fctr, sum, value.var=".n"))
# myprint_df(ctgry_cast_df)
# write.table(ctgry_cast_df, paste0(glb_out_pfx, "ctgry_cast.csv"),
# row.names=FALSE)
# print(ctgry_sum_tbl <- table(glb_allobs_df$myCategory, glb_allobs_df[, glb_rsp_var],
# useNA="ifany"))
dsp_chisq.test <- function(...) {
sel_df <- glb_allobs_df[sel_obs(...) &
!is.na(glb_allobs_df$Popular), ]
sel_df$.marker <- 1
ref_df <- glb_allobs_df[!is.na(glb_allobs_df$Popular), ]
mrg_df <- merge(ref_df[, c(glb_id_var, "Popular")],
sel_df[, c(glb_id_var, ".marker")], all.x=TRUE)
mrg_df[is.na(mrg_df)] <- 0
print(mrg_tbl <- table(mrg_df$.marker, mrg_df$Popular))
print("Rows:Selected; Cols:Popular")
#print(mrg_tbl)
print(chisq.test(mrg_tbl))
}
# dsp_chisq.test(Headline.contains="[Ee]bola")
# dsp_chisq.test(Snippet.contains="[Ee]bola")
# dsp_chisq.test(Abstract.contains="[Ee]bola")
# print(mycreate_sqlxtab_df(glb_allobs_df[sel_obs(Headline.contains="[Ee]bola"), ],
# c(glb_rsp_var, "NewsDesk", "SectionName", "SubsectionName")))
# print(table(glb_allobs_df$NewsDesk, glb_allobs_df$SectionName))
# print(table(glb_allobs_df$SectionName, glb_allobs_df$SubsectionName))
# print(table(glb_allobs_df$NewsDesk, glb_allobs_df$SectionName, glb_allobs_df$SubsectionName))
# glb_allobs_df$myCategory.fctr <- as.factor(glb_allobs_df$myCategory)
print(table(glb_allobs_df$cellular, glb_allobs_df$carrier, useNA="ifany"))##
## AT&T None Other Sprint T-Mobile Unknown Verizon
## 0 0 1586 0 0 0 0 0
## 1 288 0 3 36 28 172 196
## Unknown 4 4 2 0 0 329 0# glb_allobs_df[(glb_allobs_df$cellular %in% c("Unknown")) &
# (glb_allobs_df$carrier %in% c("AT&T", "Other")),
# c(glb_id_var, glb_rsp_var_raw, "description", "carrier", "cellular")]
glb_allobs_df[(glb_allobs_df$cellular %in% c("Unknown")) &
(glb_allobs_df$carrier %in% c("AT&T", "Other")),
"cellular"] <- "1"
# glb_allobs_df[(glb_allobs_df$cellular %in% c("Unknown")) &
# (glb_allobs_df$carrier %in% c("None")),
# c(glb_id_var, glb_rsp_var_raw, "description", "carrier", "cellular")]
glb_allobs_df[(glb_allobs_df$cellular %in% c("Unknown")) &
(glb_allobs_df$carrier %in% c("None")),
"cellular"] <- "0"
print(table(glb_allobs_df$cellular, glb_allobs_df$carrier, useNA="ifany"))##
## AT&T None Other Sprint T-Mobile Unknown Verizon
## 0 0 1590 0 0 0 0 0
## 1 292 0 5 36 28 172 196
## Unknown 0 0 0 0 0 329 0Step 2.1: scrub data
glb_chunks_df <- myadd_chunk(glb_chunks_df, "transform.data", major.inc=FALSE)## label step_major step_minor label_minor bgn end elapsed
## 3 scrub.data 2 1 1 45.357 46.438 1.081
## 4 transform.data 2 2 2 46.439 NA NA### Mapping dictionary
#sav_allobs_df <- glb_allobs_df; glb_allobs_df <- sav_allobs_df
if (!is.null(glb_map_vars)) {
for (feat in glb_map_vars) {
map_df <- myimport_data(url=glb_map_urls[[feat]],
comment="map_df",
print_diagn=TRUE)
glb_allobs_df <- mymap_codes(glb_allobs_df, feat, names(map_df)[2],
map_df, map_join_col_name=names(map_df)[1],
map_tgt_col_name=names(map_df)[2])
}
glbFeatsExclude <- union(glbFeatsExclude, glb_map_vars)
}
### Forced Assignments
#stop(here"); sav_allobs_df <- glb_allobs_df; glb_allobs_df <- sav_allobs_df
for (feat in glb_assign_vars) {
new_feat <- paste0(feat, ".my")
print(sprintf("Forced Assignments for: %s -> %s...", feat, new_feat))
glb_allobs_df[, new_feat] <- glb_allobs_df[, feat]
pairs <- glb_assign_pairs_lst[[feat]]
for (pair_ix in 1:length(pairs$from)) {
if (is.na(pairs$from[pair_ix]))
nobs <- nrow(filter(glb_allobs_df,
is.na(eval(parse(text=feat),
envir=glb_allobs_df)))) else
nobs <- sum(glb_allobs_df[, feat] == pairs$from[pair_ix])
#nobs <- nrow(filter(glb_allobs_df, is.na(Married.fctr))) ; print(nobs)
if ((is.na(pairs$from[pair_ix])) && (is.na(pairs$to[pair_ix])))
stop("what are you trying to do ???")
if (is.na(pairs$from[pair_ix]))
glb_allobs_df[is.na(glb_allobs_df[, feat]), new_feat] <-
pairs$to[pair_ix] else
glb_allobs_df[glb_allobs_df[, feat] == pairs$from[pair_ix], new_feat] <-
pairs$to[pair_ix]
print(sprintf(" %s -> %s for %s obs",
pairs$from[pair_ix], pairs$to[pair_ix], format(nobs, big.mark=",")))
}
glbFeatsExclude <- union(glbFeatsExclude, glb_assign_vars)
}
### Derivations using mapping functions
#stop(here"); sav_allobs_df <- glb_allobs_df; glb_allobs_df <- sav_allobs_df
for (new_feat in glb_derive_vars) {
print(sprintf("Creating new feature: %s...", new_feat))
args_lst <- NULL
for (arg in glbFeatsDerive[[new_feat]]$args)
args_lst[[arg]] <- glb_allobs_df[, arg]
glb_allobs_df[, new_feat] <- do.call(glbFeatsDerive[[new_feat]]$mapfn, args_lst)
}## [1] "Creating new feature: sprice.root2..."
## [1] "Creating new feature: sprice.log10..."
## [1] "Creating new feature: sprice.d20nexp..."
## [1] "Creating new feature: sprice.predict.diff..."
## [1] "Creating new feature: spdiff.cut.fctr..."
## [1] "Creating new feature: descr.my..."
## [1] "Creating new feature: prdl.my.fctr..."#stop(here")
#hex_vctr <- c("\n", "\211", "\235", "\317", "\333")
hex_regex <- paste0(c("\n", "\211", "\235", "\317", "\333"), collapse="|")
for (obs_id in c(10029, 10948, 10136, 10178, 11514, 11904, 12157, 12210, 12659)) {
# tmp_str <- unlist(strsplit(glb_allobs_df[row_pos, "descr.my"], ""))
# glb_allobs_df[row_pos, "descr.my"] <- paste0(tmp_str[!tmp_str %in% hex_vctr],
# collapse="")
row_pos <- which(glb_allobs_df$UniqueID == obs_id)
# glb_allobs_df[row_pos, "descr.my"] <-
# gsub(hex_regex, " ", glb_allobs_df[row_pos, "descr.my"])
}Step 2.2: transform data
#```{r extract_features, cache=FALSE, eval=!is.null(glbFeatsText)}
glb_chunks_df <- myadd_chunk(glb_chunks_df, "extract.features", major.inc=TRUE)## label step_major step_minor label_minor bgn end elapsed
## 4 transform.data 2 2 2 46.439 47.049 0.61
## 5 extract.features 3 0 0 47.050 NA NAextract.features_chunk_df <- myadd_chunk(NULL, "extract.features_bgn")## label step_major step_minor label_minor bgn end
## 1 extract.features_bgn 1 0 0 47.057 NA
## elapsed
## 1 NA# Create new features that help prediction
# <col_name>.lag.2 <- lag(zoo(glb_trnobs_df$<col_name>), -2, na.pad=TRUE)
# glb_trnobs_df[, "<col_name>.lag.2"] <- coredata(<col_name>.lag.2)
# <col_name>.lag.2 <- lag(zoo(glb_newobs_df$<col_name>), -2, na.pad=TRUE)
# glb_newobs_df[, "<col_name>.lag.2"] <- coredata(<col_name>.lag.2)
#
# glb_newobs_df[1, "<col_name>.lag.2"] <- glb_trnobs_df[nrow(glb_trnobs_df) - 1,
# "<col_name>"]
# glb_newobs_df[2, "<col_name>.lag.2"] <- glb_trnobs_df[nrow(glb_trnobs_df),
# "<col_name>"]
# glb_allobs_df <- mutate(glb_allobs_df,
# A.P.http=ifelse(grepl("http",Added,fixed=TRUE), 1, 0)
# )
#
# glb_trnobs_df <- mutate(glb_trnobs_df,
# )
#
# glb_newobs_df <- mutate(glb_newobs_df,
# )
# Convert dates to numbers
# typically, dates come in as chars;
# so this must be done before converting chars to factors
#stop(here"); sav_allobs_df <- glb_allobs_df #; glb_allobs_df <- sav_allobs_df
if (!is.null(glb_date_vars)) {
glb_allobs_df <- cbind(glb_allobs_df,
myextract_dates_df(df=glb_allobs_df, vars=glb_date_vars,
id_vars=glb_id_var, rsp_var=glb_rsp_var))
for (sfx in c("", ".POSIX"))
glbFeatsExclude <-
union(glbFeatsExclude,
paste(glb_date_vars, sfx, sep=""))
for (feat in glb_date_vars) {
glb_allobs_df <- orderBy(reformulate(paste0(feat, ".POSIX")), glb_allobs_df)
# print(myplot_scatter(glb_allobs_df, xcol_name=paste0(feat, ".POSIX"),
# ycol_name=glb_rsp_var, colorcol_name=glb_rsp_var))
print(myplot_scatter(glb_allobs_df[glb_allobs_df[, paste0(feat, ".POSIX")] >=
strptime("2012-12-01", "%Y-%m-%d"), ],
xcol_name=paste0(feat, ".POSIX"),
ycol_name=glb_rsp_var, colorcol_name=paste0(feat, ".wkend")))
# Create features that measure the gap between previous timestamp in the data
require(zoo)
z <- zoo(as.numeric(as.POSIXlt(glb_allobs_df[, paste0(feat, ".POSIX")])))
glb_allobs_df[, paste0(feat, ".zoo")] <- z
print(head(glb_allobs_df[, c(glb_id_var, feat, paste0(feat, ".zoo"))]))
print(myplot_scatter(glb_allobs_df[glb_allobs_df[, paste0(feat, ".POSIX")] >
strptime("2012-10-01", "%Y-%m-%d"), ],
xcol_name=paste0(feat, ".zoo"), ycol_name=glb_rsp_var,
colorcol_name=glb_rsp_var))
b <- zoo(, seq(nrow(glb_allobs_df)))
last1 <- as.numeric(merge(z-lag(z, -1), b, all=TRUE)); last1[is.na(last1)] <- 0
glb_allobs_df[, paste0(feat, ".last1.log")] <- log(1 + last1)
print(gp <- myplot_box(df=glb_allobs_df[glb_allobs_df[,
paste0(feat, ".last1.log")] > 0, ],
ycol_names=paste0(feat, ".last1.log"),
xcol_name=glb_rsp_var))
last2 <- as.numeric(merge(z-lag(z, -2), b, all=TRUE)); last2[is.na(last2)] <- 0
glb_allobs_df[, paste0(feat, ".last2.log")] <- log(1 + last2)
print(gp <- myplot_box(df=glb_allobs_df[glb_allobs_df[,
paste0(feat, ".last2.log")] > 0, ],
ycol_names=paste0(feat, ".last2.log"),
xcol_name=glb_rsp_var))
last10 <- as.numeric(merge(z-lag(z, -10), b, all=TRUE)); last10[is.na(last10)] <- 0
glb_allobs_df[, paste0(feat, ".last10.log")] <- log(1 + last10)
print(gp <- myplot_box(df=glb_allobs_df[glb_allobs_df[,
paste0(feat, ".last10.log")] > 0, ],
ycol_names=paste0(feat, ".last10.log"),
xcol_name=glb_rsp_var))
last100 <- as.numeric(merge(z-lag(z, -100), b, all=TRUE)); last100[is.na(last100)] <- 0
glb_allobs_df[, paste0(feat, ".last100.log")] <- log(1 + last100)
print(gp <- myplot_box(df=glb_allobs_df[glb_allobs_df[,
paste0(feat, ".last100.log")] > 0, ],
ycol_names=paste0(feat, ".last100.log"),
xcol_name=glb_rsp_var))
glb_allobs_df <- orderBy(reformulate(glb_id_var), glb_allobs_df)
glbFeatsExclude <- union(glbFeatsExclude,
c(paste0(feat, ".zoo")))
# all2$last3 = as.numeric(merge(z-lag(z, -3), b, all = TRUE))
# all2$last5 = as.numeric(merge(z-lag(z, -5), b, all = TRUE))
# all2$last10 = as.numeric(merge(z-lag(z, -10), b, all = TRUE))
# all2$last20 = as.numeric(merge(z-lag(z, -20), b, all = TRUE))
# all2$last50 = as.numeric(merge(z-lag(z, -50), b, all = TRUE))
#
#
# # order table
# all2 = all2[order(all2$id),]
#
# ## fill in NAs
# # count averages
# na.avg = all2 %>% group_by(weekend, hour) %>% dplyr::summarise(
# last1=mean(last1, na.rm=TRUE),
# last3=mean(last3, na.rm=TRUE),
# last5=mean(last5, na.rm=TRUE),
# last10=mean(last10, na.rm=TRUE),
# last20=mean(last20, na.rm=TRUE),
# last50=mean(last50, na.rm=TRUE)
# )
#
# # fill in averages
# na.merge = merge(all2, na.avg, by=c("weekend","hour"))
# na.merge = na.merge[order(na.merge$id),]
# for(i in c("last1", "last3", "last5", "last10", "last20", "last50")) {
# y = paste0(i, ".y")
# idx = is.na(all2[[i]])
# all2[idx,][[i]] <- na.merge[idx,][[y]]
# }
# rm(na.avg, na.merge, b, i, idx, n, pd, sec, sh, y, z)
}
}
rm(last1, last10, last100)## Warning in rm(last1, last10, last100): object 'last1' not found## Warning in rm(last1, last10, last100): object 'last10' not found## Warning in rm(last1, last10, last100): object 'last100' not found# Create factors of string variables
extract.features_chunk_df <- myadd_chunk(extract.features_chunk_df,
paste0("extract.features_", "factorize.str.vars"), major.inc=TRUE)## label step_major step_minor label_minor
## 1 extract.features_bgn 1 0 0
## 2 extract.features_factorize.str.vars 2 0 0
## bgn end elapsed
## 1 47.057 47.074 0.017
## 2 47.074 NA NA#stop(here"); sav_allobs_df <- glb_allobs_df; #glb_allobs_df <- sav_allobs_df
print(str_vars <- myfind_chr_cols_df(glb_allobs_df))## description condition cellular carrier color
## "description" "condition" "cellular" "carrier" "color"
## storage productline .src .grpid descr.my
## "storage" "productline" ".src" ".grpid" "descr.my"if (length(str_vars <- setdiff(str_vars,
c(glbFeatsExclude, glbFeatsText))) > 0) {
for (var in str_vars) {
warning("Creating factors of string variable: ", var,
": # of unique values: ", length(unique(glb_allobs_df[, var])))
glb_allobs_df[, paste0(var, ".fctr")] <-
relevel(factor(glb_allobs_df[, var]),
names(which.max(table(glb_allobs_df[, var], useNA = "ifany"))))
}
glbFeatsExclude <- union(glbFeatsExclude, str_vars)
}## Warning: Creating factors of string variable: condition: # of unique
## values: 6## Warning: Creating factors of string variable: cellular: # of unique values:
## 3## Warning: Creating factors of string variable: carrier: # of unique values:
## 7## Warning: Creating factors of string variable: color: # of unique values: 5## Warning: Creating factors of string variable: storage: # of unique values:
## 5if (!is.null(glbFeatsText)) {
require(foreach)
require(gsubfn)
require(stringr)
require(tm)
extract.features_chunk_df <- myadd_chunk(extract.features_chunk_df,
paste0("extract.features_", "process.text"), major.inc=TRUE)
chk_pattern_freq <- function(rex_str, ignore.case=TRUE) {
match_mtrx <- str_extract_all(txt_vctr, regex(rex_str, ignore_case=ignore.case),
simplify=TRUE)
match_df <- as.data.frame(match_mtrx[match_mtrx != ""])
names(match_df) <- "pattern"
return(mycreate_sqlxtab_df(match_df, "pattern"))
}
# match_lst <- gregexpr("\\bok(?!ay)", txt_vctr[746], ignore.case = FALSE, perl=TRUE); print(match_lst)
dsp_pattern <- function(rex_str, ignore.case=TRUE, print.all=TRUE) {
match_lst <- gregexpr(rex_str, txt_vctr, ignore.case = ignore.case, perl=TRUE)
match_lst <- regmatches(txt_vctr, match_lst)
match_df <- data.frame(matches=sapply(match_lst,
function (elems) paste(elems, collapse="#")))
match_df <- subset(match_df, matches != "")
if (print.all)
print(match_df)
return(match_df)
}
dsp_matches <- function(rex_str, ix) {
print(match_pos <- gregexpr(rex_str, txt_vctr[ix], perl=TRUE))
print(str_sub(txt_vctr[ix], (match_pos[[1]] / 100) * 99 + 0,
(match_pos[[1]] / 100) * 100 + 100))
}
myapply_gsub <- function(...) {
if ((length_lst <- length(names(gsub_map_lst))) == 0)
return(txt_vctr)
for (ptn_ix in 1:length_lst) {
if ((ptn_ix %% 10) == 0)
print(sprintf("running gsub for %02d (of %02d): #%s#...", ptn_ix,
length(names(gsub_map_lst)), names(gsub_map_lst)[ptn_ix]))
txt_vctr <- gsub(names(gsub_map_lst)[ptn_ix], gsub_map_lst[[ptn_ix]],
txt_vctr, ...)
}
return(txt_vctr)
}
myapply_txtmap <- function(txt_vctr, ...) {
nrows <- nrow(glb_txt_map_df)
for (ptn_ix in 1:nrows) {
if ((ptn_ix %% 10) == 0)
print(sprintf("running gsub for %02d (of %02d): #%s#...", ptn_ix,
nrows, glb_txt_map_df[ptn_ix, "rex_str"]))
txt_vctr <- gsub(glb_txt_map_df[ptn_ix, "rex_str"],
glb_txt_map_df[ptn_ix, "rpl_str"],
txt_vctr, ...)
}
return(txt_vctr)
#print(txt_vctr <- glb_allobs_df[glb_allobs_df$UniqueID == 11329, "descr.my"])
#strsplit(txt_vctr, "")[[1]][1]
#ptn_ix <- 2; glb_txt_map_df[ptn_ix, ]
#gsub(glb_txt_map_df[ptn_ix, "rex_str"], glb_txt_map_df[ptn_ix, "rpl_str"], txt_vctr)
#print(match_lst <- gregexpr(glb_txt_map_df[ptn_ix, "rex_str"], txt_vctr))
#strsplit(glb_txt_map_df[ptn_ix, "rex_str"], "")[[1]]
}
chk.equal <- function(bgn, end) {
print(all.equal(sav_txt_lst[["Headline"]][bgn:end],
glb_txt_chr_lst[["Headline"]][bgn:end]))
}
dsp.equal <- function(bgn, end) {
print(sav_txt_lst[["Headline"]][bgn:end])
print(glb_txt_chr_lst[["Headline"]][bgn:end])
}
#sav_txt_lst <- glb_txt_chr_lst; all.equal(sav_txt_lst, glb_txt_chr_lst)
#all.equal(sav_txt_lst[["Headline"]][1:4200], glb_txt_chr_lst[["Headline"]][1:4200])
#chk.equal( 1, 100)
#dsp.equal(86, 90)
txt_map_filename <- paste0(glb_txt_munge_filenames_pfx, "map.csv")
if (!file.exists(txt_map_filename))
stop(txt_map_filename, " not found!")
glb_txt_map_df <- read.csv(txt_map_filename, comment.char="#", strip.white=TRUE)
glb_txt_chr_lst <- list();
print(sprintf("Building glb_txt_chr_lst..."))
glb_txt_chr_lst <- foreach(txt_var = glbFeatsText) %dopar% {
# for (txt_var in glbFeatsText) {
txt_vctr <- glb_allobs_df[, txt_var]
names(txt_vctr) <- glb_allobs_df[, glb_id_var]
# myapply_txtmap shd be created as a tm_map::content_transformer ?
#print(glb_txt_map_df)
#txt_var=glbFeatsText[3]; txt_vctr <- glb_txt_chr_lst[[txt_var]]
#print(rex_str <- glb_txt_map_df[3, "rex_str"])
#print(rex_str <- glb_txt_map_df[glb_txt_map_df$rex_str == "\\bWall St\\.", "rex_str"])
#print(rex_str <- glb_txt_map_df[grepl("du Pont", glb_txt_map_df$rex_str), "rex_str"])
#print(rex_str <- glb_txt_map_df[glb_txt_map_df$rpl_str == "versus", "rex_str"])
#print(tmp_vctr <- grep(rex_str, txt_vctr, value=TRUE, ignore.case=FALSE))
#ret_lst <- regexec(rex_str, txt_vctr, ignore.case=FALSE); ret_lst <- regmatches(txt_vctr, ret_lst); ret_vctr <- sapply(1:length(ret_lst), function(pos_ix) ifelse(length(ret_lst[[pos_ix]]) > 0, ret_lst[[pos_ix]], "")); print(ret_vctr <- ret_vctr[ret_vctr != ""])
#gsub(rex_str, glb_txt_map_df[glb_txt_map_df$rex_str == rex_str, "rpl_str"], tmp_vctr, ignore.case=FALSE)
#grep("Hong Hong", txt_vctr, value=TRUE)
txt_vctr <- myapply_txtmap(txt_vctr, ignore.case=FALSE)
}
names(glb_txt_chr_lst) <- glbFeatsText
for (txt_var in glbFeatsText) {
print(sprintf("Remaining OK in %s:", txt_var))
txt_vctr <- glb_txt_chr_lst[[txt_var]]
print(chk_pattern_freq(rex_str <- "(?<!(BO|HO|LO))OK(?!(E\\!|ED|IE|IN|S ))",
ignore.case=FALSE))
match_df <- dsp_pattern(rex_str, ignore.case=FALSE, print.all=FALSE)
for (row in row.names(match_df))
dsp_matches(rex_str, ix=as.numeric(row))
print(chk_pattern_freq(rex_str <- "Ok(?!(a\\.|ay|in|ra|um))", ignore.case=FALSE))
match_df <- dsp_pattern(rex_str, ignore.case=FALSE, print.all=FALSE)
for (row in row.names(match_df))
dsp_matches(rex_str, ix=as.numeric(row))
print(chk_pattern_freq(rex_str <- "(?<!( b| B| c| C| g| G| j| M| p| P| w| W| r| Z|\\(b|ar|bo|Bo|co|Co|Ew|gk|go|ho|ig|jo|kb|ke|Ke|ki|lo|Lo|mo|mt|no|No|po|ra|ro|sm|Sm|Sp|to|To))ok(?!(ay|bo|e |e\\)|e,|e\\.|eb|ed|el|en|er|es|ey|i |ie|in|it|ka|ke|ki|ly|on|oy|ra|st|u |uc|uy|yl|yo))",
ignore.case=FALSE))
match_df <- dsp_pattern(rex_str, ignore.case=FALSE, print.all=FALSE)
for (row in row.names(match_df))
dsp_matches(rex_str, ix=as.numeric(row))
}
# txt_vctr <- glb_txt_chr_lst[[glbFeatsText[1]]]
# print(chk_pattern_freq(rex_str <- "(?<!( b| c| C| p|\\(b|bo|co|lo|Lo|Sp|to|To))ok(?!(ay|e |e\\)|e,|e\\.|ed|el|en|es|ey|ie|in|on|ra))", ignore.case=FALSE))
# print(chk_pattern_freq(rex_str <- "ok(?!(ay|el|on|ra))", ignore.case=FALSE))
# dsp_pattern(rex_str, ignore.case=FALSE, print.all=FALSE)
# dsp_matches(rex_str, ix=8)
# substr(txt_vctr[86], 5613, 5620)
# substr(glb_allobs_df[301, "review"], 550, 650)
#stop(here"); sav_txt_lst <- glb_txt_chr_lst
for (txt_var in glbFeatsText) {
print(sprintf("Remaining Acronyms in %s:", txt_var))
txt_vctr <- glb_txt_chr_lst[[txt_var]]
print(chk_pattern_freq(rex_str <- "([[:upper:]]\\.( *)){2,}", ignore.case=FALSE))
# Check for names
print(subset(chk_pattern_freq(rex_str <- "(([[:upper:]]+)\\.( *)){1}",
ignore.case=FALSE),
.n > 1))
# dsp_pattern(rex_str="(OK\\.( *)){1}", ignore.case=FALSE)
# dsp_matches(rex_str="(OK\\.( *)){1}", ix=557)
#dsp_matches(rex_str="\\bR\\.I\\.P(\\.*)(\\B)", ix=461)
#dsp_matches(rex_str="\\bR\\.I\\.P(\\.*)", ix=461)
#print(str_sub(txt_vctr[676], 10100, 10200))
#print(str_sub(txt_vctr[74], 1, -1))
}
for (txt_var in glbFeatsText) {
re_str <- "\\b(Fort|Ft\\.|Hong|Las|Los|New|Puerto|Saint|San|St\\.)( |-)(\\w)+"
print(sprintf("Remaining #%s# terms in %s: ", re_str, txt_var))
txt_vctr <- glb_txt_chr_lst[[txt_var]]
print(orderBy(~ -.n +pattern, subset(chk_pattern_freq(re_str, ignore.case=FALSE),
grepl("( |-)[[:upper:]]", pattern))))
print(" consider cleaning if relevant to problem domain; geography name; .n > 1")
#grep("New G", txt_vctr, value=TRUE, ignore.case=FALSE)
#grep("St\\. Wins", txt_vctr, value=TRUE, ignore.case=FALSE)
}
#stop(here"); sav_txt_lst <- glb_txt_chr_lst
for (txt_var in glbFeatsText) {
re_str <- "\\b(N|S|E|W|C)( |\\.)(\\w)+"
print(sprintf("Remaining #%s# terms in %s: ", re_str, txt_var))
txt_vctr <- glb_txt_chr_lst[[txt_var]]
print(orderBy(~ -.n +pattern, subset(chk_pattern_freq(re_str, ignore.case=FALSE),
grepl(".", pattern))))
#grep("N Weaver", txt_vctr, value=TRUE, ignore.case=FALSE)
}
for (txt_var in glbFeatsText) {
re_str <- "\\b(North|South|East|West|Central)( |\\.)(\\w)+"
print(sprintf("Remaining #%s# terms in %s: ", re_str, txt_var))
txt_vctr <- glb_txt_chr_lst[[txt_var]]
if (nrow(filtered_df <- subset(chk_pattern_freq(re_str, ignore.case=FALSE),
grepl(".", pattern))) > 0)
print(orderBy(~ -.n +pattern, filtered_df))
#grep("Central (African|Bankers|Cast|Italy|Role|Spring)", txt_vctr, value=TRUE, ignore.case=FALSE)
#grep("East (Africa|Berlin|London|Poland|Rivals|Spring)", txt_vctr, value=TRUE, ignore.case=FALSE)
#grep("North (American|Korean|West)", txt_vctr, value=TRUE, ignore.case=FALSE)
#grep("South (Pacific|Street)", txt_vctr, value=TRUE, ignore.case=FALSE)
#grep("St\\. Martins", txt_vctr, value=TRUE, ignore.case=FALSE)
}
find_cmpnd_wrds <- function(txt_vctr) {
# Enhancements:
# - arg should be txt_corpus instead of txt_vctr
txt_corpus <- Corpus(VectorSource(txt_vctr))
txt_corpus <- tm_map(txt_corpus, content_transformer(tolower), lazy = TRUE)
txt_corpus <- tm_map(txt_corpus, PlainTextDocument, lazy = TRUE)
txt_corpus <- tm_map(txt_corpus, removePunctuation,
preserve_intra_word_dashes = TRUE, lazy = FALSE)
# Defaulting to Tf since TfIdf with normalize = TRUE throws a warning for empty docs
terms_mtrx <- as.matrix(TermDocumentMatrix(txt_corpus,
control = list(weighting = weightTf)))
terms_df <- orderBy(~ -Tf, data.frame(term = dimnames(terms_mtrx)$Terms,
Tf = rowSums(terms_mtrx)))
cmpnd_df <- subset(terms_df, grepl("-", term))
if (nrow(cmpnd_df) == 0) {
print(" No compounded terms found")
return(FALSE)
}
txt_compound_filename <- paste0(glb_txt_munge_filenames_pfx, "compound.csv")
if (!file.exists(txt_compound_filename))
stop(txt_compound_filename, " not found!")
filter_df <- read.csv(txt_compound_filename, comment.char="#", strip.white=TRUE)
cmpnd_df$filter <- FALSE
for (row_ix in 1:nrow(filter_df))
cmpnd_df[!cmpnd_df$filter, "filter"] <-
grepl(filter_df[row_ix, "rex_str"],
cmpnd_df[!cmpnd_df$filter, "term"], ignore.case=TRUE)
cmpnd_df <- subset(cmpnd_df, !filter)
# Bug in tm_map(txt_corpus, removePunctuation, preserve_intra_word_dashes=TRUE) ???
# "net-a-porter" gets converted to "net-aporter"
#grep("net-a-porter", txt_vctr, ignore.case=TRUE, value=TRUE)
#grep("maser-laser", txt_vctr, ignore.case=TRUE, value=TRUE)
#txt_corpus[[which(grepl("net-a-porter", txt_vctr, ignore.case=TRUE))]]
#grep("\\b(across|longer)-(\\w)", cmpnd_df$term, ignore.case=TRUE, value=TRUE)
#grep("(\\w)-(affected|term)\\b", cmpnd_df$term, ignore.case=TRUE, value=TRUE)
print(sprintf("nrow(cmpnd_df): %d", nrow(cmpnd_df)))
myprint_df(cmpnd_df)
}
# This should be run after glb_txt_corpus_lst is created with tolower
extract.features_chunk_df <- myadd_chunk(extract.features_chunk_df,
paste0("extract.features_", "process.text_reporting_compound_terms"), major.inc=FALSE)
# for (txt_var in glbFeatsText) {
# print(sprintf("Remaining compound terms in %s: ", txt_var))
# find_cmpnd_wrds(txt_vctr = glb_txt_chr_lst[[txt_var]])
# #grep("thirty-five", txt_vctr, ignore.case=TRUE, value=TRUE)
# #rex_str <- glb_txt_map_df[grepl("hirty", glb_txt_map_df$rex_str), "rex_str"]
# }
extract.features_chunk_df <- myadd_chunk(extract.features_chunk_df,
paste0("extract.features_", "build.corpus"), major.inc=TRUE)
get_txt_terms <- function(terms_TDM) {
terms_mtrx <- as.matrix(as.TermDocumentMatrix(terms_TDM))
docms_mtrx <- as.matrix(as.DocumentTermMatrix(terms_TDM))
terms_df <- data.frame(term = dimnames(terms_mtrx)$Terms,
weight = rowSums(terms_mtrx),
freq = rowSums(terms_mtrx > 0))
terms_df$pos <- 1:nrow(terms_df)
terms_df$cor.y <-
cor(docms_mtrx[glb_allobs_df$.src == "Train",],
as.numeric(glb_allobs_df[glb_allobs_df$.src == "Train", glb_rsp_var]),
use = "pairwise.complete.obs")
terms_df$cor.y.abs <- abs(terms_df$cor.y)
# .rnorm.cor.y.abs <- abs(cor(glb_allobs_df[glb_allobs_df$.src == "Train", ".rnorm"],
# as.numeric(glb_allobs_df[glb_allobs_df$.src == "Train", glb_rsp_var]),
# use = "pairwise.complete.obs"))
terms_df$chisq.stat <- NA; terms_df$chisq.pval <- NA
for (ix in 1:nrow(terms_df)) {
#for (ix in 1:743) {
if (length(unique(docms_mtrx[glb_allobs_df$.src == "Train", ix])) > 1) {
chisq <- chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
glb_allobs_df[glb_allobs_df$.src == "Train", glb_rsp_var])
terms_df[ix, "chisq.stat"] <- chisq$statistic
terms_df[ix, "chisq.pval"] <- chisq$p.value
}
}
nzv_df <- nzv(docms_mtrx[glb_allobs_df$.src == "Train",], freqCut = glb_nzv_freqCut,
uniqueCut = glb_nzv_uniqueCut, saveMetrics = TRUE)
terms_df$nzv.freqRatio <- nzv_df$freqRatio
# terms_df$nzv.freqRatio.cut.fctr <- cut(terms_df$nzv.freqRatio,
# breaks = sort(c(min(terms_df$nzv.freqRatio),
# glb_nzv_freqCut,
# max(terms_df$nzv.freqRatio))))
terms_df$nzv.percentUnique <- nzv_df$percentUnique
# terms_df$nzv.percentUnique.cut.fctr <- cut(terms_df$nzv.percentUnique,
# breaks = sort(c(min(terms_df$nzv.percentUnique) - .Machine$double.neg.eps,
# glb_nzv_uniqueCut,
# max(terms_df$nzv.percentUnique))))
# terms_df$nzv.quad.fctr <- as.factor(paste0("fRatio:", terms_df$nzv.freqRatio.cut.fctr,
# "\n%Unq:", terms_df$nzv.percentUnique.cut.fctr))
terms_df$nzv <- nzv_df$nzv
for (cls in unique(glb_allobs_df[, glb_txt_cor_var])) {
if (!is.na(cls))
terms_df[, paste0("weight.", as.character(cls))] <-
colSums(t(terms_mtrx) *
as.numeric(!is.na(glb_allobs_df[, glb_txt_cor_var]) &
(glb_allobs_df[, glb_txt_cor_var] == cls))) else
terms_df[, paste0("weight.", as.character(cls))] <-
colSums(t(terms_mtrx) *
as.numeric(is.na(glb_allobs_df[, glb_txt_cor_var])))
}
# Check all calls to get_terms_DTM_terms to change returned order assumption
return(terms_df <- orderBy(~ -weight, terms_df))
}
#plt_full_df <- get_terms_DTM_terms(terms_DTM=glb_full_terms_DTM_lst[[txt_var]])
get_corpus_terms <- function(txt_corpus) {
return(terms_df <- get_txt_terms(terms_TDM =
TermDocumentMatrix(txt_corpus, control = glb_txt_terms_control)))
}
myreplacePunctuation <- function(x) {
return(gsub("[[:punct:]]+", " ", gsub("('|-)", "", x)))
}
#stop(here"); glb_to_sav()
glb_txt_corpus_lst <- list()
print(sprintf("Building glb_txt_corpus_lst..."))
glb_txt_corpus_lst <- foreach(txt_var = glbFeatsText, .verbose = FALSE) %dopar% {
#glb_txt_corpus_lst <- foreach(txt_var = glbFeatsText, .verbose = TRUE) %do% {
#for (txt_var in glbFeatsText) {
txt_corpus <- Corpus(VectorSource(glb_txt_chr_lst[[txt_var]]))
txt_corpus <- tm_map(txt_corpus, PlainTextDocument, lazy = TRUE)
txt_corpus <- tm_map(txt_corpus, content_transformer(tolower), lazy = TRUE) #nuppr
# removePunctuation does not replace with whitespace. Use a custom transformer ???
#txt_corpus <- tm_map(txt_corpus, removePunctuation, lazy = TRUE) #npnct<chr_ix>
txt_corpus <- tm_map(txt_corpus, content_transformer(myreplacePunctuation)
, lazy = TRUE) #npnct<chr_ix>
# txt-corpus <- tm_map(txt_corpus, content_transformer(function(x, pattern) gsub(pattern, "", x))
if (!is.null(glb_txt_stop_words[[txt_var]]))
txt_corpus <- tm_map(txt_corpus, removeWords, glb_txt_stop_words[[txt_var]],
lazy = FALSE)#, lazy=TRUE) #nstopwrds
# foreach result is based on .Last.Eval
txt_corpus <- txt_corpus
# glb_txt_corpus_lst[[txt_var]] <- txt_corpus
}
names(glb_txt_corpus_lst) <- glbFeatsText
mycombineSynonyms <- content_transformer(function(x, syn=NULL) {
Reduce(function(a,b) {
gsub(paste0("\\b(", paste(b$syns, collapse = "|"),")\\b"), b$word, a)}, syn, x)
})
#stop(here"); glb_to_sav(); sav_txt_corpus <- glb_txt_corpus_lst[[txt_var]]; all.equal(sav_txt_corpus, glb_txt_corpus_lst[[txt_var]]); glb_txt_corpus_lst[[txt_var]] <- sav_txt_corpus
glb_post_stop_words_terms_df_lst <- list();
glb_post_stop_words_terms_mtrx_lst <- list();
glb_post_stem_words_terms_df_lst <- list();
glb_post_stem_words_terms_mtrx_lst <- list();
for (txt_var in glbFeatsText) {
print(sprintf(" Top_n post-stop term weights for %s:", txt_var))
# This impacts stemming probably due to lazy parameter
print(myprint_df(full_terms_df <-
get_corpus_terms(txt_corpus=glb_txt_corpus_lst[[txt_var]]),
glbFeatsTextTermsMax[[txt_var]]))
glb_post_stop_words_terms_df_lst[[txt_var]] <- full_terms_df
terms_stop_mtrx <- as.matrix(DocumentTermMatrix(glb_txt_corpus_lst[[txt_var]],
control=glb_txt_terms_control))
rownames(terms_stop_mtrx) <- rownames(glb_allobs_df) # print undreadable otherwise
glb_post_stop_words_terms_mtrx_lst[[txt_var]] <- terms_stop_mtrx
tmp_allobs_df <- glb_allobs_df[, c(glb_id_var, glb_rsp_var)]
tmp_allobs_df$terms.post.stop.n <- rowSums(terms_stop_mtrx > 0)
tmp_allobs_df$terms.post.stop.n.log <- log(1 + tmp_allobs_df$terms.post.stop.n)
tmp_allobs_df$weight.post.stop.sum <- rowSums(terms_stop_mtrx)
print(sprintf(" Top_n stem term weights for %s:", txt_var))
glb_txt_corpus_lst[[txt_var]] <- tm_map(glb_txt_corpus_lst[[txt_var]], stemDocument,
"english", lazy=FALSE)
if (!is.null(glb_txt_synonyms[[txt_var]])) {
syn_lst <- myrmNullObj(glb_txt_synonyms[[txt_var]])
glb_txt_corpus_lst[[txt_var]] <- tm_map(glb_txt_corpus_lst[[txt_var]],
mycombineSynonyms,
syn_lst, lazy=FALSE)
}
print(myprint_df(full_terms_df <- get_corpus_terms(glb_txt_corpus_lst[[txt_var]]),
glbFeatsTextTermsMax[[txt_var]]))
glb_post_stem_words_terms_df_lst[[txt_var]] <- full_terms_df
terms_stem_mtrx <- as.matrix(DocumentTermMatrix(glb_txt_corpus_lst[[txt_var]],
control=glb_txt_terms_control))
rownames(terms_stem_mtrx) <- rownames(glb_allobs_df) # print undreadable otherwise
glb_post_stem_words_terms_mtrx_lst[[txt_var]] <- terms_stem_mtrx
tmp_allobs_df$terms.post.stem.n <- rowSums(terms_stem_mtrx > 0)
tmp_allobs_df$terms.post.stem.n.log <- log(1 + tmp_allobs_df$terms.post.stem.n)
tmp_allobs_df$weight.post.stem.sum <- rowSums(terms_stem_mtrx)
tmp_allobs_df$terms.n.stem.stop.Ratio <-
1.0 * tmp_allobs_df$terms.post.stem.n / tmp_allobs_df$terms.post.stop.n
tmp_allobs_df[(is.nan(tmp_allobs_df$terms.n.stem.stop.Ratio) |
is.infinite(tmp_allobs_df$terms.n.stem.stop.Ratio)),
"terms.n.stem.stop.Ratio"] <- 1.0
if ((n.errors <- sum(tmp_allobs_df$terms.n.stem.stop.Ratio > 1)) > 0)
stop(n.errors, " obs in tmp_allobs_df have terms.n.stem.stop.Ratio > 1",
" happening due to terms filtered by glb_txt_terms_control$bounds$global[1] but stemmable to other terms")
#print(head(subset(tmp_allobs_df, terms.n.stem.stop.Ratio > 1)))
#glb_allobs_df[(row_ix <- which(glb_allobs_df$UniqueID == 10465)), ]
#terms_stop_mtrx[row_ix, terms_stop_mtrx[row_ix, ] > 0]
#setdiff(names(terms_stem_mtrx[row_ix, terms_stem_mtrx[row_ix, ] > 0]), names(terms_stop_mtrx[row_ix, terms_stop_mtrx[row_ix, ] > 0]))
#mydsp_obs(list(descr.my.contains="updat"))
tmp_allobs_df$weight.sum.stem.stop.Ratio <-
1.0 * tmp_allobs_df$weight.post.stem.sum / tmp_allobs_df$weight.post.stop.sum
tmp_allobs_df[is.nan(tmp_allobs_df$weight.sum.stem.stop.Ratio) |
is.infinite(tmp_allobs_df$weight.sum.stem.stop.Ratio),
"weight.sum.stem.stop.Ratio"] <- 1.0
tmp_trnobs_df <- tmp_allobs_df[!is.na(tmp_allobs_df[, glb_rsp_var]), ]
print(cor(as.matrix(tmp_trnobs_df[, -c(1, 2)]),
as.numeric(tmp_trnobs_df[, glb_rsp_var])))
txt_var_pfx <- toupper(substr(txt_var, 1, 1))
tmp_allobs_df <- tmp_allobs_df[, -c(1, 2)]
names(tmp_allobs_df) <- paste(paste0(txt_var_pfx, "."), names(tmp_allobs_df), sep = "")
glb_allobs_df <- cbind(glb_allobs_df, tmp_allobs_df)
glbFeatsExclude <- c(glbFeatsExclude,
paste(paste0(txt_var_pfx, ".terms.post."), c("stop.n", "stem.n"), sep = ""))
}
require(wordcloud)
for (txt_var in glbFeatsText) {
print(sprintf(" Wordcloud post-stem term weights for %s:", txt_var))
m <- glb_post_stem_words_terms_mtrx_lst[[txt_var]]
# calculate the frequency of words
v <- sort(colSums(m), decreasing = TRUE)
myNames <- names(v)
d <- data.frame(word = myNames, freq = v)
wordcloud(d$word, d$freq, min.freq = glb_txt_terms_control$bounds$global[1])
}
for (txt_var in glbFeatsText) {
.rnorm.cor.y.abs <- abs(cor(glb_allobs_df[glb_allobs_df$.src == "Train", ".rnorm"],
as.numeric(glb_allobs_df[glb_allobs_df$.src == "Train", glb_rsp_var]),
use = "pairwise.complete.obs"))
plt_df <- subset(glb_post_stem_words_terms_df_lst[[txt_var]], !is.na(cor.y))
plt_df$nzv.freqRatio.cut.fctr <- cut(plt_df$nzv.freqRatio,
breaks = sort(c(min(plt_df$nzv.freqRatio),
glb_nzv_freqCut,
max(plt_df$nzv.freqRatio))))
plt_df$nzv.percentUnique.cut.fctr <- cut(plt_df$nzv.percentUnique,
breaks = sort(c(min(plt_df$nzv.percentUnique) - .Machine$double.neg.eps,
glb_nzv_uniqueCut,
max(plt_df$nzv.percentUnique))))
plt_df$nzv.quad.fctr <- as.factor(paste0("fRatio:", plt_df$nzv.freqRatio.cut.fctr,
"\n%Unq:", plt_df$nzv.percentUnique.cut.fctr))
labelCnd <- !plt_df$nzv |
(!is.na(plt_df$chisq.pval) & (plt_df$chisq.pval < 0.05)) &
(!is.na(plt_df$cor.y.abs) & (plt_df$cor.y.abs > .rnorm.cor.y.abs))
plt_df$label <- NA; plt_df[labelCnd, "label"] <- plt_df[labelCnd, "term"]
print(ggplot(plt_df, aes(x = cor.y, y = chisq.stat)) +
geom_point(aes(color = nzv.quad.fctr, size = weight)) +
geom_text(aes(label = label), color = "gray50") +
# geom_vline(xintercept = 0) +
geom_vline(xintercept = c(-1, +1) * .rnorm.cor.y.abs, color = "gray",
linetype = "dashed") +
ggtitle(txt_var))
}
extract.features_chunk_df <- myadd_chunk(extract.features_chunk_df,
paste0("extract.features_", "extract.DTM"), major.inc=TRUE)
#stop(here")
glb_full_DTM_lst <- list(); glb_sprs_DTM_lst <- list();
for (txt_var in glbFeatsText) {
print(sprintf("Extracting term weights for %s...", txt_var))
txt_corpus <- glb_txt_corpus_lst[[txt_var]]
full_DTM <- DocumentTermMatrix(txt_corpus,
control=glb_txt_terms_control)
sprs_DTM <- removeSparseTerms(full_DTM,
glb_sprs_thresholds[txt_var])
glb_full_DTM_lst[[txt_var]] <- full_DTM
glb_sprs_DTM_lst[[txt_var]] <- sprs_DTM
}
extract.features_chunk_df <- myadd_chunk(extract.features_chunk_df,
paste0("extract.features_", "report.DTM"), major.inc=TRUE)
require(reshape2)
for (txt_var in glbFeatsText) {
print(sprintf("Reporting term weights for %s...", txt_var))
full_DTM <- glb_full_DTM_lst[[txt_var]]
sprs_DTM <- glb_sprs_DTM_lst[[txt_var]]
print(" Full TermMatrix:"); print(full_DTM)
full_terms_df <- get_txt_terms(full_DTM)
# full_terms_df <- full_terms_df[, c(2, 1, 3, 4)]
# col_names <- names(full_terms_df)
# col_names[2:length(col_names)] <-
# paste(col_names[2:length(col_names)], ".full", sep="")
# names(full_terms_df) <- col_names
print(" Sparse TermMatrix:"); print(sprs_DTM)
sprs_terms_df <- get_txt_terms(sprs_DTM)
# sprs_terms_df <- sprs_terms_df[, c(2, 1, 3, 4)]
# col_names <- names(sprs_terms_df)
# col_names[2:length(col_names)] <-
# paste(col_names[2:length(col_names)], ".sprs", sep="")
# names(sprs_terms_df) <- col_names
#intersect(names(full_terms_df), names(sprs_terms_df))
terms_df <- merge(full_terms_df, sprs_terms_df, by = c("term", "weight", "freq",
grep("weight\\.", names(full_terms_df), value = TRUE)),
all.x = TRUE, suffixes = c(".full", ".sprs"))
terms_df$in.sprs <- !is.na(terms_df$pos.sprs)
plt_terms_df <- subset(terms_df,
weight >= min(terms_df$weight[!is.na(terms_df$pos.sprs)], na.rm=TRUE))
plt_terms_df$label <- ""
plt_terms_df[is.na(plt_terms_df$pos.sprs), "label"] <-
plt_terms_df[is.na(plt_terms_df$pos.sprs), "term"]
# glb_important_terms[[txt_var]] <- union(glb_important_terms[[txt_var]],
# plt_terms_df[is.na(plt_terms_df$TfIdf.sprs), "term"])
print(myplot_scatter(plt_terms_df, "freq", "weight",
colorcol_name="in.sprs") +
geom_text(aes(label=label), color="Black", size=3.5))
melt_terms_df <- orderBy(~ -value,
melt(terms_df, id.vars="term", measure.vars = c("weight", "freq")))
print(ggplot(melt_terms_df, aes(value, color=variable)) + stat_ecdf() +
geom_hline(yintercept=glb_sprs_thresholds[txt_var],
linetype = "dotted"))
melt_terms_df <- orderBy(~ -value,
melt(subset(terms_df, in.sprs), id.vars="term",
measure.vars=grep("weight.", names(terms_df), value=TRUE)))
print(myplot_hbar(melt_terms_df, "term", "value", colorcol_name="variable"))
melt_terms_df <- orderBy(~ -value,
melt(subset(terms_df, !in.sprs), id.vars="term",
measure.vars=grep("weight.", names(terms_df), value=TRUE)))
print(myplot_hbar(head(melt_terms_df, glbFeatsTextTermsMax[[txt_var]]), "term", "value",
colorcol_name="variable"))
}
# sav_full_DTM_lst <- glb_full_DTM_lst
# print(identical(sav_glb_txt_corpus_lst, glb_txt_corpus_lst))
# print(all.equal(length(sav_glb_txt_corpus_lst), length(glb_txt_corpus_lst)))
# print(all.equal(names(sav_glb_txt_corpus_lst), names(glb_txt_corpus_lst)))
# print(all.equal(sav_glb_txt_corpus_lst[["Headline"]], glb_txt_corpus_lst[["Headline"]]))
# print(identical(sav_full_DTM_lst, glb_full_DTM_lst))
rm(full_terms_mtrx)
# Create txt features
if ((length(glbFeatsText) > 1) &&
(length(unique(pfxs <- sapply(glbFeatsText,
function(txt) toupper(substr(txt, 1, 1))))) < length(glbFeatsText)))
stop("Prefixes for corpus freq terms not unique: ", pfxs)
extract.features_chunk_df <- myadd_chunk(extract.features_chunk_df,
paste0("extract.features_", "bind.DTM"),
major.inc=TRUE)
#stop(here"); glb_to_sav(); all.equal(sav_allobs_df, glb_allobs_df); glb_allobs_df <- sav_allobs_df
require(tidyr)
for (txt_var in glbFeatsText) {
print(sprintf("Binding DTM for %s...", txt_var))
txt_var_pfx <- toupper(substr(txt_var, 1, 1))
txt_full_X_df <- as.data.frame(as.matrix(glb_full_DTM_lst[[txt_var]]))
terms_full_df <- get_txt_terms(glb_full_DTM_lst[[txt_var]])
# make.names adds a period to R keywords e.g. "in", "function"
colnames(txt_full_X_df) <- paste(txt_var_pfx, ".T.",
make.names(colnames(txt_full_X_df)), sep="")
rownames(txt_full_X_df) <- rownames(glb_allobs_df) # warning otherwise
# plt_full_df <- terms_full_df
# names(plt_full_df)[grepl("weight$", names(plt_full_df))] <- "weight.all"
# # gather(plt_full_df[1:5, ], domain, TfIdf, -matches("!(TfIdf)"))
# # gather(plt_full_df[1:5, grepl("TfIdf", names(plt_full_df))], domain, TfIdf)
# # gather(plt_full_df[1:5, ], domain, TfIdf,
# # -names(plt_full_df)[!grepl("TfIdf", names(plt_full_df))])
# plt_full_df <- gather(plt_full_df, domain, weight,
# -c(term, freq, pos, cor.y, cor.y.abs))
# plt_full_df$label <- NA
# top_val_terms <- orderBy(~-weight, terms_full_df)$term[1:glbFeatsTextTermsMax[[txt_var]]]
# plt_full_df[plt_full_df$term %in% top_val_terms, "label"] <-
# plt_full_df[plt_full_df$term %in% top_val_terms, "term"]
# top_cor_terms <- orderBy(~-cor.y.abs,
# terms_full_df)$term[1:glbFeatsTextTermsMax[[txt_var]]]
# plt_full_df[plt_full_df$term %in% top_cor_terms, "label"] <-
# plt_full_df[plt_full_df$term %in% top_cor_terms, "term"]
# #plt_full_df$type <- "none"
# plt_full_df[plt_full_df$term %in% top_val_terms, "type"] <- "top.weight"
# plt_full_df[plt_full_df$term %in% top_cor_terms, "type"] <- "top.cor"
# plt_full_df[plt_full_df$term %in% intersect(top_val_terms, top_cor_terms), "type"] <-
# "top.both"
# cor.y.rnorm <- cor(glb_allobs_df$.rnorm, as.numeric(glb_allobs_df[, glb_rsp_var]),
# use = "pairwise.complete.obs")
# print(ggplot(plt_full_df, aes(x=weight, y=cor.y)) + facet_wrap(~ domain) +
# geom_point(aes(size=freq), color="grey") +
# geom_jitter() +
# geom_text(aes(label=label, color=type), size=3.5) +
# #geom_hline(yintercept=cor.y.rnorm, color="red") +
# geom_hline(yintercept=c(cor.y.rnorm, -cor.y.rnorm), color="red"))
#stop(here"); glb_to_sav()
if (glbFeatsTextFilter == "sparse") {
txt_X_df <- as.data.frame(as.matrix(glb_sprs_DTM_lst[[txt_var]]))
select_terms <- make.names(colnames(txt_X_df))
# colnames(txt_X_df) <- paste(txt_var_pfx, ".T.",
# make.names(colnames(txt_X_df)), sep="")
# rownames(txt_X_df) <- rownames(glb_allobs_df) # warning otherwise
} else if (glbFeatsTextFilter == "top.val") {
select_terms <- orderBy(~-weight,
terms_full_df)$term[1:glbFeatsTextTermsMax[[txt_var]]]
# txt_X_df <- txt_full_X_df[, subset(terms_full_df, term %in% select_terms)$pos,
# FALSE]
} else if (glbFeatsTextFilter == "top.cor") {
select_terms <- orderBy(~-cor.y.abs,
terms_full_df)$term[1:glbFeatsTextTermsMax[[txt_var]]]
# txt_X_df <- txt_full_X_df[, subset(terms_full_df, term %in% select_terms)$pos,
# FALSE]
} else if (glbFeatsTextFilter == "top.chisq") {
select_terms <- orderBy(~-chisq.stat,
subset(terms_full_df, chisq.pval < 0.05)
)$term[1:glbFeatsTextTermsMax[[txt_var]]]
} else if (glbFeatsTextFilter == "union.top.val.cor") {
select_terms <- union(
orderBy(~-weight , terms_full_df)$term[1:glbFeatsTextTermsMax[[txt_var]]],
orderBy(~-cor.y.abs, terms_full_df)$term[1:glbFeatsTextTermsMax[[txt_var]]])
} else stop(
"glbFeatsTextFilter should be one of c('sparse', 'top.val', 'top.cor', 'union.top.val.cor', 'top.chisq') vs. '",
glbFeatsTextFilter, "'")
assoc_terms_lst <- findAssocs(glb_full_DTM_lst[[txt_var]], select_terms,
glbFeatsTextAssocCor[[txt_var]])
assoc_terms <- c(NULL)
for (term in names(assoc_terms_lst))
if (length(assoc_terms_lst[[term]]) > 0)
assoc_terms <- union(assoc_terms, names(assoc_terms_lst[[term]]))
#stop(here"); glb_to_sav()
txt_X_df <- txt_full_X_df[,
subset(terms_full_df, term %in% c(select_terms, assoc_terms))$pos,
FALSE]
glb_allobs_df <- cbind(glb_allobs_df, txt_X_df) # TfIdf is normalized
#glb_allobs_df <- cbind(glb_allobs_df, log_X_df) # if using non-normalized metrics
}
#identical(chk_entity_df, glb_allobs_df)
#chk_entity_df <- glb_allobs_df
extract.features_chunk_df <- myadd_chunk(extract.features_chunk_df,
paste0("extract.features_", "bind.DXM"),
major.inc=TRUE)
#stop(here"); sav_allobs_df <- glb_allobs_df; glb_allobs_df <- sav_allobs_df
glb_punct_vctr <- c("!", "\"", "#", "\\$", "%", "&", "'",
"\\(|\\)",# "\\(", "\\)",
"\\*", "\\+", ",", "-", "\\.", "/", ":", ";",
"<|>", # "<",
"=",
# ">",
"\\?", "@", "\\[", "\\\\", "\\]", "\\^", "_", "`",
"\\{", "\\|", "\\}", "~")
txt_X_df <- glb_allobs_df[, c(glb_id_var, ".rnorm"), FALSE]
txt_X_df <- foreach(txt_var=glbFeatsText, .combine=cbind) %dopar% {
#for (txt_var in glbFeatsText) {
print(sprintf("Binding DXM for %s...", txt_var))
txt_var_pfx <- toupper(substr(txt_var, 1, 1))
txt_full_DTM_mtrx <- as.matrix(glb_full_DTM_lst[[txt_var]])
rownames(txt_full_DTM_mtrx) <- rownames(glb_allobs_df) # print undreadable otherwise
#print(txt_full_DTM_mtrx[txt_full_DTM_mtrx[, "ebola"] != 0, "ebola"])
# Create <txt_var>.T.<term> for glb_important_terms
for (term in glb_important_terms[[txt_var]])
txt_X_df[, paste0(txt_var_pfx, ".T.", make.names(term))] <-
txt_full_DTM_mtrx[, term]
# Create <txt_var>.wrds.n.log & .wrds.unq.n.log
txt_X_df[, paste0(txt_var_pfx, ".wrds.n.log")] <-
log(1 + mycount_pattern_occ("\\w+", glb_txt_chr_lst[[txt_var]]))
txt_X_df[, paste0(txt_var_pfx, ".wrds.unq.n.log")] <-
log(1 + rowSums(txt_full_DTM_mtrx != 0))
txt_X_df[, paste0(txt_var_pfx, ".weight.sum")] <-
rowSums(txt_full_DTM_mtrx)
txt_X_df[, paste0(txt_var_pfx, ".ratio.weight.sum.wrds.n")] <-
txt_X_df[, paste0(txt_var_pfx, ".weight.sum")] /
(exp(txt_X_df[, paste0(txt_var_pfx, ".wrds.n.log")]) - 1)
txt_X_df[is.nan(txt_X_df[, paste0(txt_var_pfx, ".ratio.weight.sum.wrds.n")]),
paste0(txt_var_pfx, ".ratio.weight.sum.wrds.n")] <- 0
# Create <txt_var>.chrs.n.log
txt_X_df[, paste0(txt_var_pfx, ".chrs.n.log")] <-
log(1 + mycount_pattern_occ(".", glb_allobs_df[, txt_var]))
txt_X_df[, paste0(txt_var_pfx, ".chrs.uppr.n.log")] <-
log(1 + mycount_pattern_occ("[[:upper:]]", glb_allobs_df[, txt_var]))
txt_X_df[, paste0(txt_var_pfx, ".dgts.n.log")] <-
log(1 + mycount_pattern_occ("[[:digit:]]", glb_allobs_df[, txt_var]))
# Create <txt_var>.npnct?.log
# would this be faster if it's iterated over each row instead of
# each created column ???
for (punct_ix in 1:length(glb_punct_vctr)) {
# smp0 <- " "
# smp1 <- "! \" # $ % & ' ( ) * + , - . / : ; < = > ? @ [ \ ] ^ _ ` { | } ~"
# smp2 <- paste(smp1, smp1, sep=" ")
# print(sprintf("Testing %s pattern:", glb_punct_vctr[punct_ix]))
# results <- mycount_pattern_occ(glb_punct_vctr[punct_ix], c(smp0, smp1, smp2))
# names(results) <- NULL; print(results)
txt_X_df[,
paste0(txt_var_pfx, ".chrs.pnct", sprintf("%02d", punct_ix), ".n.log")] <-
log(1 + mycount_pattern_occ(glb_punct_vctr[punct_ix],
glb_allobs_df[, txt_var]))
}
# print(head(glb_allobs_df[glb_allobs_df[, "A.npnct23.log"] > 0,
# c("UniqueID", "Popular", "Abstract", "A.npnct23.log")]))
# Create <txt_var>.wrds.stop.n.log & <txt_var>ratio.wrds.stop.n.wrds.n
if (!is.null(glb_txt_stop_words[[txt_var]])) {
stop_words_rex_str <- paste0("\\b(",
paste0(glb_txt_stop_words[[txt_var]], collapse="|"),
")\\b")
txt_X_df[, paste0(txt_var_pfx, ".wrds.stop.n", ".log")] <-
log(1 + mycount_pattern_occ(stop_words_rex_str, glb_txt_chr_lst[[txt_var]]))
txt_X_df[, paste0(txt_var_pfx, ".ratio.wrds.stop.n.wrds.n")] <-
exp(txt_X_df[, paste0(txt_var_pfx, ".wrds.stop.n", ".log")] -
txt_X_df[, paste0(txt_var_pfx, ".wrds.n", ".log")])
}
# Create <txt_var>.P.http
txt_X_df[, paste(txt_var_pfx, ".P.http", sep="")] <-
as.integer(0 + mycount_pattern_occ("http", glb_allobs_df[, txt_var]))
# Create <txt_var>.P.mini & air
txt_X_df[, paste(txt_var_pfx, ".P.mini", sep="")] <-
as.integer(0 + mycount_pattern_occ("mini(?!m)", glb_allobs_df[, txt_var],
perl=TRUE))
txt_X_df[, paste(txt_var_pfx, ".P.air", sep="")] <-
as.integer(0 + mycount_pattern_occ("(?<![fhp])air", glb_allobs_df[, txt_var],
perl=TRUE))
txt_X_df[, paste(txt_var_pfx, ".P.black", sep="")] <-
as.integer(0 + mycount_pattern_occ("black", glb_allobs_df[, txt_var],
perl=TRUE))
txt_X_df[, paste(txt_var_pfx, ".P.white", sep="")] <-
as.integer(0 + mycount_pattern_occ("white", glb_allobs_df[, txt_var],
perl=TRUE))
txt_X_df[, paste(txt_var_pfx, ".P.gold", sep="")] <-
as.integer(0 + mycount_pattern_occ("gold", glb_allobs_df[, txt_var],
perl=TRUE))
txt_X_df[, paste(txt_var_pfx, ".P.spacegray", sep="")] <-
as.integer(0 + mycount_pattern_occ("spacegray", glb_allobs_df[, txt_var],
perl=TRUE))
txt_X_df <- subset(txt_X_df, select=-.rnorm)
txt_X_df <- txt_X_df[, -grep(glb_id_var, names(txt_X_df), fixed=TRUE), FALSE]
#glb_allobs_df <- cbind(glb_allobs_df, txt_X_df)
}
glb_allobs_df <- cbind(glb_allobs_df, txt_X_df)
#myplot_box(glb_allobs_df, "A.sum.TfIdf", glb_rsp_var)
# if (sum(is.na(glb_allobs_df$D.P.http)) > 0)
# stop("Why is this happening ?")
# Generate summaries
# print(summary(glb_allobs_df))
# print(sapply(names(glb_allobs_df), function(col) sum(is.na(glb_allobs_df[, col]))))
# print(summary(glb_trnobs_df))
# print(sapply(names(glb_trnobs_df), function(col) sum(is.na(glb_trnobs_df[, col]))))
# print(summary(glb_newobs_df))
# print(sapply(names(glb_newobs_df), function(col) sum(is.na(glb_newobs_df[, col]))))
glbFeatsExclude <- union(glbFeatsExclude,
glbFeatsText)
rm(log_X_df, txt_X_df)
}## Loading required package: stringr## label step_major step_minor label_minor
## 2 extract.features_factorize.str.vars 2 0 0
## 3 extract.features_process.text 3 0 0
## bgn end elapsed
## 2 47.074 47.144 0.071
## 3 47.145 NA NA
## [1] "Building glb_txt_chr_lst..."
## [1] "running gsub for 10 (of 179): #\\bCentral African Republic\\b#..."
## [1] "running gsub for 20 (of 179): #\\bAlejandro G\\. Iñárritu#..."
## [1] "running gsub for 30 (of 179): #\\bC\\.A\\.A\\.#..."
## [1] "running gsub for 40 (of 179): #\\bCV\\.#..."
## [1] "running gsub for 50 (of 179): #\\bE\\.P\\.A\\.#..."
## [1] "running gsub for 60 (of 179): #\\bG\\.I\\. Joe#..."
## [1] "running gsub for 70 (of 179): #\\bISIS\\.#..."
## [1] "running gsub for 80 (of 179): #\\bJ\\.K\\. Simmons#..."
## [1] "running gsub for 90 (of 179): #\\bM\\. Henri Pol#..."
## [1] "running gsub for 100 (of 179): #\\bN\\.Y\\.S\\.E\\.#..."
## [1] "running gsub for 110 (of 179): #\\bR\\.B\\.S\\.#..."
## [1] "running gsub for 120 (of 179): #\\bSteven A\\. Cohen#..."
## [1] "running gsub for 130 (of 179): #\\bV\\.A\\.#..."
## [1] "running gsub for 140 (of 179): #\\bWall Street#..."
## [1] "running gsub for 150 (of 179): #\\bSaint( |-)((Laurent|Lucia)\\b)+#..."
## [1] "running gsub for 160 (of 179): #\\bSouth( |\\\\.)(America|American|Africa|African|Carolina|Dakota|Korea|Korean|Sudan)\\b#..."
## [1] "running gsub for 170 (of 179): #(\\w)-a-year#..."
## [1] "Remaining OK in descr.my:"
## pattern .n
## 1 OK 6
## [[1]]
## [1] NA
## attr(,"match.length")
## [1] NA
##
## [1] NA
## [[1]]
## [1] NA
## attr(,"match.length")
## [1] NA
##
## [1] NA
## [[1]]
## [1] NA
## attr(,"match.length")
## [1] NA
##
## [1] NA
## [[1]]
## [1] NA
## attr(,"match.length")
## [1] NA
##
## [1] NA
## [[1]]
## [1] NA
## attr(,"match.length")
## [1] NA
##
## [1] NA
## [[1]]
## [1] NA
## attr(,"match.length")
## [1] NA
##
## [1] NA
## [1] pattern .n
## <0 rows> (or 0-length row.names)
## [1] pattern .n
## <0 rows> (or 0-length row.names)
## [1] "Remaining Acronyms in descr.my:"
## [1] pattern .n
## <0 rows> (or 0-length row.names)
## pattern .n
## 1 CONDITION. 6
## 2 ONLY. 5
## 3 GB. 4
## 4 BOX. 2
## 5 CORNER. 2
## 6 ESN. 2
## 7 GOOD. 2
## 8 ICLOUD. 2
## 9 IMEI. 2
## 10 IPADS. 2
## 11 LOCKED. 2
## 12 LOCKS. 2
## 13 ON. 2
## 14 SCRATCHES. 2
## 15 TEARS. 2
## 16 USE. 2
## 17 WIFIONLY. 2
## [1] "Remaining #\\b(Fort|Ft\\.|Hong|Las|Los|New|Puerto|Saint|San|St\\.)( |-)(\\w)+# terms in descr.my: "
## pattern .n
## 1 New Open 3
## 4 New Digitizer 1
## 5 New Opened 1
## 6 New Screen 1
## [1] " consider cleaning if relevant to problem domain; geography name; .n > 1"
## [1] "Remaining #\\b(N|S|E|W|C)( |\\.)(\\w)+# terms in descr.my: "
## pattern .n
## 1 C Stock 3
## 2 W blue 1
## [1] "Remaining #\\b(North|South|East|West|Central)( |\\.)(\\w)+# terms in descr.my: "
## label step_major
## 3 extract.features_process.text 3
## 4 extract.features_process.text_reporting_compound_terms 3
## step_minor label_minor bgn end elapsed
## 3 0 0 47.145 48.734 1.589
## 4 1 1 48.735 NA NA
## label step_major
## 4 extract.features_process.text_reporting_compound_terms 3
## 5 extract.features_build.corpus 4
## step_minor label_minor bgn end elapsed
## 4 1 1 48.735 48.74 0.005
## 5 0 0 48.741 NA NA
## [1] "Building glb_txt_corpus_lst..."
## [1] " Top_n post-stop term weights for descr.my:"## Warning in weighting(x): empty document(s): character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) charact## Warning in cor(docms_mtrx[glb_allobs_df$.src == "Train", ],
## as.numeric(glb_allobs_df[glb_allobs_df$.src == : the standard deviation is
## zero## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## [1] "Rows: 347; Cols: 14"
## term weight freq pos cor.y cor.y.abs chisq.stat
## condition condition 160.57765 486 69 -0.034976682 0.034976682 69.49175
## likenew likenew 123.78010 70 175 -0.043168766 0.043168766 22.07653
## in in 102.43546 432 146 -0.071653919 0.071653919 40.30764
## used used 96.66889 229 327 0.008747051 0.008747051 31.92471
## and and 92.37461 331 21 0.006010357 0.006010357 38.45444
## is is 87.72451 335 162 -0.042931491 0.042931491 26.34374
## chisq.pval nzv.freqRatio nzv.percentUnique nzv weight.N
## condition 2.609830e-06 36.97561 1.351351 TRUE 61.61577
## likenew 1.817939e-01 257.71429 0.972973 TRUE 54.11495
## in 1.981535e-02 33.52174 1.351351 TRUE 46.12932
## used 1.017193e-01 54.54839 1.297297 TRUE 33.84672
## and 7.095553e-02 55.24138 1.513514 TRUE 35.73653
## is 5.541564e-01 54.06667 1.567568 TRUE 35.26992
## weight.Y weight.NA
## condition 42.65757 56.30431
## likenew 19.52427 50.14088
## in 26.82731 29.47884
## used 31.76561 31.05656
## and 31.78712 24.85096
## is 23.45595 28.99864
## term weight freq pos cor.y cor.y.abs chisq.stat
## case case 33.533127 76 54 -0.009225083 0.009225083 16.106811
## protector protector 10.206604 18 253 0.006965022 0.006965022 8.315515
## contact contact 10.006409 15 71 0.035032399 0.035032399 9.951124
## 64gb 64gb 8.260734 14 10 -0.030822325 0.030822325 8.352751
## otterbox otterbox 7.546463 10 221 -0.013289954 0.013289954 4.917015
## look look 1.267528 2 182 -0.030370111 0.030370111 1.720453
## chisq.pval nzv.freqRatio nzv.percentUnique nzv weight.N
## case 0.5850931 198.4444 1.0270270 TRUE 16.019773
## protector 0.5026891 611.6667 0.5405405 TRUE 4.336370
## contact 0.0412591 262.5714 0.2702703 TRUE 2.440087
## 64gb 0.3025184 613.6667 0.4324324 TRUE 4.043603
## otterbox 0.5544999 921.5000 0.3783784 TRUE 3.401559
## look 0.4230663 1848.0000 0.1621622 TRUE 1.267528
## weight.Y weight.NA
## case 12.327119 5.186234
## protector 4.386232 1.484003
## contact 5.843907 1.722415
## 64gb 1.260555 2.956575
## otterbox 1.826941 2.317964
## look 0.000000 0.000000
## term weight freq pos cor.y cor.y.abs
## having having 0.7106680 1 140 NA NA
## shipped shipped 0.7106680 1 284 0.02508765 0.02508765
## thats thats 0.6688640 1 310 -0.02155773 0.02155773
## gentle gentle 0.6317049 1 130 -0.02155773 0.02155773
## opening opening 0.6317049 1 216 0.02508765 0.02508765
## previously previously 0.5984572 1 246 NA NA
## chisq.stat chisq.pval nzv.freqRatio nzv.percentUnique nzv
## having NA NA 0 0.05405405 TRUE
## shipped 5.762908e-03 0.9394877 1849 0.10810811 TRUE
## thats 1.281078e-27 1.0000000 1849 0.10810811 TRUE
## gentle 1.281078e-27 1.0000000 1849 0.10810811 TRUE
## opening 5.762908e-03 0.9394877 1849 0.10810811 TRUE
## previously NA NA 0 0.05405405 TRUE
## weight.N weight.Y weight.NA
## having 0.0000000 0.0000000 0.7106680
## shipped 0.0000000 0.7106680 0.0000000
## thats 0.6688640 0.0000000 0.0000000
## gentle 0.6317049 0.0000000 0.0000000
## opening 0.0000000 0.6317049 0.0000000
## previously 0.0000000 0.0000000 0.5984572
## term weight freq pos cor.y cor.y.abs
## having having 0.7106680 1 140 NA NA
## shipped shipped 0.7106680 1 284 0.02508765 0.02508765
## thats thats 0.6688640 1 310 -0.02155773 0.02155773
## gentle gentle 0.6317049 1 130 -0.02155773 0.02155773
## opening opening 0.6317049 1 216 0.02508765 0.02508765
## previously previously 0.5984572 1 246 NA NA
## chisq.stat chisq.pval nzv.freqRatio nzv.percentUnique nzv
## having NA NA 0 0.05405405 TRUE
## shipped 5.762908e-03 0.9394877 1849 0.10810811 TRUE
## thats 1.281078e-27 1.0000000 1849 0.10810811 TRUE
## gentle 1.281078e-27 1.0000000 1849 0.10810811 TRUE
## opening 5.762908e-03 0.9394877 1849 0.10810811 TRUE
## previously NA NA 0 0.05405405 TRUE
## weight.N weight.Y weight.NA
## having 0.0000000 0.0000000 0.7106680
## shipped 0.0000000 0.7106680 0.0000000
## thats 0.6688640 0.0000000 0.0000000
## gentle 0.6317049 0.0000000 0.0000000
## opening 0.0000000 0.6317049 0.0000000
## previously 0.0000000 0.0000000 0.5984572## Warning in weighting(x): empty document(s): character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) charact## [1] " Top_n stem term weights for descr.my:"## Warning in weighting(x): empty document(s): character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) charact## Warning in cor(docms_mtrx[glb_allobs_df$.src == "Train", ],
## as.numeric(glb_allobs_df[glb_allobs_df$.src == : the standard deviation is
## zero## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## [1] "Rows: 226; Cols: 14"
## term weight freq pos cor.y cor.y.abs chisq.stat
## condit condit 160.55353 492 52 -0.033248924 0.033248924 68.76178
## likenew likenew 123.78010 70 117 -0.043168766 0.043168766 22.07653
## use use 111.06226 285 210 0.003025811 0.003025811 45.28121
## in in 102.43546 432 98 -0.071653919 0.071653919 40.30764
## and and 92.37461 331 20 0.006010357 0.006010357 38.45444
## is is 87.72451 335 110 -0.042931491 0.042931491 26.34374
## chisq.pval nzv.freqRatio nzv.percentUnique nzv weight.N
## condit 3.361980e-06 36.82927 1.351351 TRUE 61.51671
## likenew 1.817939e-01 257.71429 0.972973 TRUE 54.11495
## use 1.518411e-02 86.63158 1.513514 TRUE 40.77891
## in 1.981535e-02 33.52174 1.351351 TRUE 46.12932
## and 7.095553e-02 55.24138 1.513514 TRUE 35.73653
## is 5.541564e-01 54.06667 1.567568 TRUE 35.26992
## weight.Y weight.NA
## condit 43.14002 55.89681
## likenew 19.52427 50.14088
## use 35.93377 34.34958
## in 26.82731 29.47884
## and 31.78712 24.85096
## is 23.45595 28.99864
## term weight freq pos cor.y cor.y.abs chisq.stat
## brandnew brandnew 37.154496 30 36 -0.014982381 0.014982381 13.498032
## doesnt doesnt 25.357692 57 67 -0.052798870 0.052798870 20.355151
## never never 11.829767 15 132 -0.003422891 0.003422891 7.655336
## set set 10.427103 13 178 0.035657781 0.035657781 8.724937
## restor restor 9.331353 13 168 -0.005404144 0.005404144 9.811595
## absolut absolut 8.451809 12 12 0.066008565 0.066008565 9.350376
## chisq.pval nzv.freqRatio nzv.percentUnique nzv weight.N
## brandnew 0.56388835 609.0000 0.8648649 TRUE 21.081563
## doesnt 0.06065895 129.4286 0.7027027 TRUE 11.924374
## never 0.36396438 459.7500 0.4324324 TRUE 4.776183
## set 0.18964858 920.0000 0.3783784 TRUE 2.246840
## restor 0.19950403 613.3333 0.4324324 TRUE 4.584716
## absolut 0.15481081 921.0000 0.3783784 TRUE 0.000000
## weight.Y weight.NA
## brandnew 11.651061 4.421872
## doesnt 4.248782 9.184536
## never 3.700799 3.352785
## set 5.534968 2.645295
## restor 3.357971 1.388666
## absolut 6.251752 2.200057
## term weight freq pos cor.y cor.y.abs chisq.stat
## 4g 4g 2.200200 3 9 0.0002965679 0.0002965679 2.023052e+00
## sprint sprint 2.037099 2 191 -0.0215577316 0.0215577316 1.281078e-27
## verizon verizon 2.037099 2 212 -0.0215577316 0.0215577316 1.281078e-27
## tmobil tmobil 1.893062 2 203 0.0094480109 0.0094480109 2.023052e+00
## ipad1 ipad1 1.834814 2 101 0.0067428569 0.0067428569 2.023052e+00
## gold gold 1.137069 1 86 -0.0215577316 0.0215577316 1.281078e-27
## chisq.pval nzv.freqRatio nzv.percentUnique nzv weight.N
## 4g 0.3636637 1848 0.1621622 TRUE 0.6989804
## sprint 1.0000000 1849 0.1081081 TRUE 0.7407634
## verizon 1.0000000 1849 0.1081081 TRUE 0.7407634
## tmobil 0.3636637 1848 0.1621622 TRUE 0.7407634
## ipad1 0.3636637 1848 0.1621622 TRUE 0.7977452
## gold 1.0000000 1849 0.1081081 TRUE 1.1370687
## weight.Y weight.NA
## 4g 0.6116078 0.8896114
## sprint 0.0000000 1.2963359
## verizon 0.0000000 1.2963359
## tmobil 1.1522986 0.0000000
## ipad1 1.0370687 0.0000000
## gold 0.0000000 0.0000000
## term weight freq pos cor.y cor.y.abs chisq.stat
## 4g 4g 2.200200 3 9 0.0002965679 0.0002965679 2.023052e+00
## sprint sprint 2.037099 2 191 -0.0215577316 0.0215577316 1.281078e-27
## verizon verizon 2.037099 2 212 -0.0215577316 0.0215577316 1.281078e-27
## tmobil tmobil 1.893062 2 203 0.0094480109 0.0094480109 2.023052e+00
## ipad1 ipad1 1.834814 2 101 0.0067428569 0.0067428569 2.023052e+00
## gold gold 1.137069 1 86 -0.0215577316 0.0215577316 1.281078e-27
## chisq.pval nzv.freqRatio nzv.percentUnique nzv weight.N
## 4g 0.3636637 1848 0.1621622 TRUE 0.6989804
## sprint 1.0000000 1849 0.1081081 TRUE 0.7407634
## verizon 1.0000000 1849 0.1081081 TRUE 0.7407634
## tmobil 0.3636637 1848 0.1621622 TRUE 0.7407634
## ipad1 0.3636637 1848 0.1621622 TRUE 0.7977452
## gold 1.0000000 1849 0.1081081 TRUE 1.1370687
## weight.Y weight.NA
## 4g 0.6116078 0.8896114
## sprint 0.0000000 1.2963359
## verizon 0.0000000 1.2963359
## tmobil 1.1522986 0.0000000
## ipad1 1.0370687 0.0000000
## gold 0.0000000 0.0000000## Warning in weighting(x): empty document(s): character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) charact## [,1]
## terms.post.stop.n -0.075731613
## terms.post.stop.n.log -0.062526435
## weight.post.stop.sum -0.045009208
## terms.post.stem.n -0.074990216
## terms.post.stem.n.log -0.062220631
## weight.post.stem.sum -0.047105442
## terms.n.stem.stop.Ratio 0.043562113
## weight.sum.stem.stop.Ratio 0.002381527## Loading required package: wordcloud
## Loading required package: RColorBrewer## [1] " Wordcloud post-stem term weights for descr.my:"## Warning in wordcloud(d$word, d$freq, min.freq = glb_txt_terms_control
## $bounds$global[1]): likenew could not be fit on page. It will not be
## plotted.
## Warning: Removed 195 rows containing missing values (geom_text).## label step_major step_minor label_minor bgn
## 5 extract.features_build.corpus 4 0 0 48.741
## 6 extract.features_extract.DTM 5 0 0 71.465
## end elapsed
## 5 71.464 22.723
## 6 NA NA
## [1] "Extracting term weights for descr.my..."## Warning in weighting(x): empty document(s): character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) character(0) character(0) character(0) character(0)
## character(0) charact## label step_major step_minor label_minor bgn
## 6 extract.features_extract.DTM 5 0 0 71.465
## 7 extract.features_report.DTM 6 0 0 72.831
## end elapsed
## 6 72.831 1.366
## 7 NA NA
## [1] "Reporting term weights for descr.my..."
## [1] " Full TermMatrix:"
## <<DocumentTermMatrix (documents: 2648, terms: 226)>>
## Non-/sparse entries: 11341/587107
## Sparsity : 98%
## Maximal term length: 10
## Weighting : term frequency - inverse document frequency (normalized) (tf-idf)## Warning in cor(docms_mtrx[glb_allobs_df$.src == "Train", ],
## as.numeric(glb_allobs_df[glb_allobs_df$.src == : the standard deviation is
## zero## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## [1] " Sparse TermMatrix:"
## <<DocumentTermMatrix (documents: 2648, terms: 19)>>
## Non-/sparse entries: 4701/45611
## Sparsity : 91%
## Maximal term length: 7
## Weighting : term frequency - inverse document frequency (normalized) (tf-idf)## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in myplot_scatter(plt_terms_df, "freq", "weight", colorcol_name =
## "in.sprs"): converting in.sprs to class:factor
## Warning in rm(full_terms_mtrx): object 'full_terms_mtrx' not found## label step_major step_minor label_minor bgn
## 7 extract.features_report.DTM 6 0 0 72.831
## 8 extract.features_bind.DTM 7 0 0 77.562
## end elapsed
## 7 77.561 4.731
## 8 NA NA## Loading required package: tidyr## [1] "Binding DTM for descr.my..."## Warning in cor(docms_mtrx[glb_allobs_df$.src == "Train", ],
## as.numeric(glb_allobs_df[glb_allobs_df$.src == : the standard deviation is
## zero## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## Warning in chisq.test(docms_mtrx[glb_allobs_df$.src == "Train", ix],
## glb_allobs_df[glb_allobs_df$.src == : Chi-squared approximation may be
## incorrect## label step_major step_minor label_minor bgn
## 8 extract.features_bind.DTM 7 0 0 77.562
## 9 extract.features_bind.DXM 8 0 0 80.382
## end elapsed
## 8 80.382 2.82
## 9 NA NA
## [1] "Binding DXM for descr.my..."## Warning in rm(log_X_df, txt_X_df): object 'log_X_df' not found
# Use model info provided in description
# mydsp_obs(list(description.contains="a[[:digit:]]"), cols=glb_dsp_cols, all=TRUE)
# glb_allobs_df[glb_allobs_df$UniqueID == 12474, "prdline.my"] <- "iPad mini"
# glb_allobs_df[glb_allobs_df$UniqueID == 12474, "color"] <- "Space Gray"
# glb_allobs_df[glb_allobs_df$UniqueID == 12474, "cellular"] <- "0"
# glb_allobs_df[glb_allobs_df$UniqueID == 12474, "carrier"] <- "None"
#
# mydsp_obs(list(description.contains="m(.{4})ll"), cols=glb_dsp_cols, all=TRUE)
# glb_allobs_df[glb_allobs_df$UniqueID == 11360, "color"] <- "Black"
# glb_allobs_df[glb_allobs_df$UniqueID == 11360, "storage"] <- "64"
# glb_allobs_df[glb_allobs_df$UniqueID == 11360, "cellular"] <- "0"
# glb_allobs_df[glb_allobs_df$UniqueID == 11360, "carrier"] <- "None"
#
# glb_allobs_df[glb_allobs_df$UniqueID == 11361, "prdline.my"] <- "iPad Air"
# glb_allobs_df[glb_allobs_df$UniqueID == 11361, "storage"] <- "32"
# glb_allobs_df[glb_allobs_df$UniqueID == 11361, "color"] <- "White"
# glb_allobs_df[glb_allobs_df$UniqueID == 11361, "cellular"] <- "0"
# glb_allobs_df[glb_allobs_df$UniqueID == 11361, "carrier"] <- "None"
# mydsp_obs(list(description.contains="mini(?!m)"), perl=TRUE, cols="D.P.mini", all=TRUE)
# mydsp_obs(list(D.P.mini=1), cols="D.P.mini", all=TRUE)
# mydsp_obs(list(D.P.mini=1, productline="Unknown"), cols="D.P.mini", all=TRUE)
# mydsp_obs(list(description.contains="(?<![fhp])air"), perl=TRUE, all=TRUE)
# mydsp_obs(list(description.contains="air"), perl=FALSE, cols="D.P.air", all=TRUE)
# mydsp_obs(list(D.P.air=1, productline="Unknown"), cols="D.P.air", all=TRUE)
# print(mycreate_sqlxtab_df(glb_allobs_df, c("prdline.my", "productline", "D.P.mini",
# glb_rsp_var)))
# print(glb_allobs_df[(glb_allobs_df$productline == "Unknown") &
# (glb_allobs_df$D.P.mini > 0),
# c(glb_id_var, glb_category_var, glb_dsp_cols, glbFeatsText)])
# glb_allobs_df[(glb_allobs_df$D.P.mini == 1) & (glb_allobs_df$productline == "Unknown"),
# "prdline.my"] <- "iPad mini"
# print(mycreate_sqlxtab_df(glb_allobs_df, c("prdline.my", "productline", "D.P.air",
# glb_rsp_var)))
# print(glb_allobs_df[(glb_allobs_df$productline == "Unknown") &
# (glb_allobs_df$D.P.air > 0),
# c(glb_id_var, glb_category_var, glb_dsp_cols, glbFeatsText)])
# #glb_allobs_df[glb_allobs_df$UniqueID == 11863, "D.P.air"] <- 0
# glb_allobs_df[(glb_allobs_df$D.P.air == 1) & (glb_allobs_df$productline == "Unknown"),
# "prdline.my"] <- "iPad Air"
# print(glb_allobs_df[(glb_allobs_df$UniqueID %in% c(11767, 11811, 12156)),
# c(glb_id_var, "sold",
# "prdline.my", "color", "condition", "cellular", "carrier", "storage"
# #, "descr.my"
# )])
# glb_allobs_df[glb_allobs_df$UniqueID == 11767, "prdline.my"] <- "iPad 2"
# glb_allobs_df[glb_allobs_df$UniqueID == 11767, "storage"] <- "32"
# glb_allobs_df[glb_allobs_df$UniqueID == 11811, "prdline.my"] <- "iPad 2"
# glb_allobs_df[glb_allobs_df$UniqueID == 12156, "prdline.my"] <- "iPad 1"
# mydsp_obs(list(prdline.my="Unknown"), all=TRUE)
# tmp_allobs_df <- glb_allobs_df[, "prdline.my", FALSE]
# names(tmp_allobs_df) <- "old.prdline.my"
# glb_allobs_df$prdline.my <-
# plyr::revalue(glb_allobs_df$prdline.my, c(
# # "iPad 1" = "iPad",
# # "iPad 2" = "iPad2+",
# "iPad 3" = "iPad 3+",
# "iPad 4" = "iPad 3+",
# "iPad 5" = "iPad 3+",
#
# "iPad Air" = "iPadAir",
# "iPad Air 2" = "iPadAir",
#
# "iPad mini" = "iPadmini",
# "iPad mini 2" = "iPadmini 2+",
# "iPad mini 3" = "iPadmini 2+",
# "iPad mini Retina" = "iPadmini 2+"
# ))
# tmp_allobs_df$prdline.my <- glb_allobs_df[, "prdline.my"]
# print(mycreate_sqlxtab_df(tmp_allobs_df, c("prdline.my", "old.prdline.my")))
# print(mycreate_sqlxtab_df(tmp_allobs_df, c("prdline.my")))
# print(mycreate_sqlxtab_df(subset(glb_allobs_df, color == "Unknown"),
# c("color", "D.P.black", "D.P.gold", "D.P.spacegray", "D.P.white")))
# print(glb_allobs_df[(glb_allobs_df$color == "Unknown") & (glb_allobs_df$D.P.black > 0),
# c(glb_id_var, "color", "D.P.black", "sold", "prdline.my", "condition",
# "cellular", "carrier", "storage", "descr.my")])
# glb_allobs_df[glb_allobs_df$UniqueID == 12137, "color"] <- "Black"
# print(glb_allobs_df[(glb_allobs_df$color == "Unknown") & (glb_allobs_df$D.P.spacegray > 0),
# c(glb_id_var, "color", "D.P.spacegray", "prdline.my", "condition",
# "cellular", "carrier", "storage", "descr.my")])
# glb_allobs_df[glb_allobs_df$UniqueID %in% c(12106), "color"] <- "Space Gray"
# print(glb_allobs_df[(glb_allobs_df$color == "Unknown") & (glb_allobs_df$D.P.white > 0),
# c(glb_id_var, "color", "D.P.white", "prdline.my", "condition",
# "cellular", "carrier", "storage", "descr.my")])
# glb_allobs_df[glb_allobs_df$UniqueID %in% c(10573, 10809, 10925, 11735), "color"] <-
# "White"
glb_allobs_df$carrier.fctr <- as.factor(glb_allobs_df$carrier)
glb_allobs_df$cellular.fctr <- as.factor(glb_allobs_df$cellular)
glb_allobs_df$color.fctr <- as.factor(glb_allobs_df$color)
# glb_allobs_df$prdline.my.fctr <- as.factor(glb_allobs_df$prdline.my)
glb_allobs_df$storage.fctr <- as.factor(glb_allobs_df$storage)
#stop(here"); sav_allobs_df <- glb_allobs_df; glb_allobs_df <- sav_allobs_df
# glb_allobs_df %>%
# unite(prdl.descr.my, c(productline, as.numeric(D.chrs.n.log > 0), sep="#"))
# unite_("prdl.descr.my", interp(~c("productline", as.numeric(D.chrs.n.log > 0), sep="#")))
# print(sapply(names(glb_trnobs_df), function(col) sum(is.na(glb_trnobs_df[, col]))))
# print(sapply(names(glb_newobs_df), function(col) sum(is.na(glb_newobs_df[, col]))))
# print(myplot_scatter(glb_trnobs_df, "<col1_name>", "<col2_name>", smooth=TRUE))
#stop(here"); glb_to_sav(); glb_allobs_df <- sav_allobs_df
if (!is.null(glbFeatsPrice)) {
for (var in glbFeatsPrice) {
for (digit in 1:(log10(max(glb_allobs_df[, var], na.rm=TRUE)) + 1)) {
glb_allobs_df[, paste0(var, ".dgt", digit, ".is9")] <-
as.numeric(as.integer((as.integer(glb_allobs_df[, var]) %% (10 ^ digit)) /
(10 ^ (digit - 1))) == 9)
# glb_allobs_df[, paste0(var, ".dgt", digit, ".is9.fctr")] <-
# as.factor(as.integer((as.integer(glb_allobs_df[, var]) %% (10 ^ digit)) /
# (10 ^ (digit - 1))) == 9)
}
for (decimal in 1:2) {
glb_allobs_df[, paste0(var, ".dcm", decimal, ".is9")] <-
as.numeric(as.integer(glb_allobs_df[, var] * (10 ^ decimal)) %% 10 == 9)
# glb_allobs_df[, paste0(var, ".dcm", decimal, ".is9.fctr")] <-
# as.factor(as.integer(glb_allobs_df[, var] * (10 ^ decimal)) %% 10 == 9)
}
}
#as.numeric((as.integer(startprice) %% 10) == 9)
}
rm(corpus_lst
, glb_sprs_DTM_lst #, glb_full_DTM_lst
, txt_corpus, txt_vctr)## Warning in rm(corpus_lst, glb_sprs_DTM_lst, txt_corpus, txt_vctr): object
## 'corpus_lst' not foundextract.features_chunk_df <- myadd_chunk(extract.features_chunk_df, "extract.features_end",
major.inc=TRUE)## label step_major step_minor label_minor bgn
## 9 extract.features_bind.DXM 8 0 0 80.382
## 10 extract.features_end 9 0 0 274.829
## end elapsed
## 9 274.829 194.447
## 10 NA NAmyplt_chunk(extract.features_chunk_df)## label step_major
## 9 extract.features_bind.DXM 8
## 5 extract.features_build.corpus 4
## 7 extract.features_report.DTM 6
## 8 extract.features_bind.DTM 7
## 3 extract.features_process.text 3
## 6 extract.features_extract.DTM 5
## 2 extract.features_factorize.str.vars 2
## 1 extract.features_bgn 1
## 4 extract.features_process.text_reporting_compound_terms 3
## step_minor label_minor bgn end elapsed duration
## 9 0 0 80.382 274.829 194.447 194.447
## 5 0 0 48.741 71.464 22.723 22.723
## 7 0 0 72.831 77.561 4.731 4.730
## 8 0 0 77.562 80.382 2.820 2.820
## 3 0 0 47.145 48.734 1.589 1.589
## 6 0 0 71.465 72.831 1.366 1.366
## 2 0 0 47.074 47.144 0.071 0.070
## 1 0 0 47.057 47.074 0.017 0.017
## 4 1 1 48.735 48.740 0.005 0.005
## [1] "Total Elapsed Time: 274.829 secs"
# if (glb_save_envir)
# save(glb_feats_df,
# glb_allobs_df, #glb_trnobs_df, glb_fitobs_df, glb_OOBobs_df, glb_newobs_df,
# file=paste0(glb_out_pfx, "extract_features_dsk.RData"))
# load(paste0(glb_out_pfx, "extract_features_dsk.RData"))
replay.petrisim(pn=glb_analytics_pn,
replay.trans=(glb_analytics_avl_objs <- c(glb_analytics_avl_objs,
"data.training.all","data.new")), flip_coord=TRUE)## time trans "bgn " "fit.data.training.all " "predict.data.new " "end "
## 0.0000 multiple enabled transitions: data.training.all data.new model.selected firing: data.training.all
## 1.0000 1 2 1 0 0
## 1.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction firing: data.new
## 2.0000 2 1 1 1 0
#glb_chunks_df <- myadd_chunk(glb_chunks_df, "manage.missing.data", major.inc=TRUE)Step 3.0: extract features
glb_chunks_df <- myadd_chunk(glb_chunks_df, "manage.missing.data", major.inc=FALSE)## label step_major step_minor label_minor bgn end
## 5 extract.features 3 0 0 47.050 280.281
## 6 manage.missing.data 3 1 1 280.282 NA
## elapsed
## 5 233.231
## 6 NA# If mice crashes with error: Error in get(as.character(FUN), mode = "function", envir = envir) : object 'State' of mode 'function' was not found
# consider excluding 'State' as a feature
# print(sapply(names(glb_trnobs_df), function(col) sum(is.na(glb_trnobs_df[, col]))))
# print(sapply(names(glb_newobs_df), function(col) sum(is.na(glb_newobs_df[, col]))))
# glb_trnobs_df <- na.omit(glb_trnobs_df)
# glb_newobs_df <- na.omit(glb_newobs_df)
# df[is.na(df)] <- 0
mycheck_problem_data(glb_allobs_df, featsExclude = glbFeatsExclude,
fctrMaxUniqVals = glbFctrMaxUniqVals)## [1] "numeric data missing in : "
## sold sold.fctr
## 798 798
## [1] "numeric data w/ 0s in : "
## biddable sold
## 1437 995
## sprice.log10 cellular.fctr
## 31 1590
## D.terms.post.stop.n D.terms.post.stop.n.log
## 1516 1516
## D.weight.post.stop.sum D.terms.post.stem.n
## 1516 1516
## D.terms.post.stem.n.log D.weight.post.stem.sum
## 1516 1516
## D.T.condit D.T.use
## 2156 2363
## D.T.in. D.T.good
## 2216 2451
## D.T.screen D.T.with
## 2440 2434
## D.T.of D.T.mint
## 2497 2585
## D.T.or D.T.cosmet
## 2523 2531
## D.T.minor D.T.light
## 2531 2567
## D.T.X100 D.T.from
## 2584 2591
## D.T.hous D.wrds.n.log
## 2576 1513
## D.wrds.unq.n.log D.weight.sum
## 1516 1516
## D.ratio.weight.sum.wrds.n D.chrs.n.log
## 1516 1513
## D.chrs.uppr.n.log D.dgts.n.log
## 1515 2452
## D.chrs.pnct01.n.log D.chrs.pnct02.n.log
## 2570 2648
## D.chrs.pnct03.n.log D.chrs.pnct04.n.log
## 2647 2648
## D.chrs.pnct05.n.log D.chrs.pnct06.n.log
## 2582 2596
## D.chrs.pnct07.n.log D.chrs.pnct08.n.log
## 2611 2572
## D.chrs.pnct09.n.log D.chrs.pnct10.n.log
## 2632 2639
## D.chrs.pnct11.n.log D.chrs.pnct12.n.log
## 2294 2534
## D.chrs.pnct13.n.log D.chrs.pnct14.n.log
## 1930 2574
## D.chrs.pnct15.n.log D.chrs.pnct16.n.log
## 2628 2639
## D.chrs.pnct17.n.log D.chrs.pnct18.n.log
## 2646 2647
## D.chrs.pnct19.n.log D.chrs.pnct20.n.log
## 2648 2648
## D.chrs.pnct21.n.log D.chrs.pnct22.n.log
## 2648 2648
## D.chrs.pnct23.n.log D.chrs.pnct24.n.log
## 2648 2648
## D.chrs.pnct25.n.log D.chrs.pnct26.n.log
## 2648 2648
## D.chrs.pnct27.n.log D.chrs.pnct28.n.log
## 2648 2640
## D.chrs.pnct29.n.log D.chrs.pnct30.n.log
## 2648 2648
## D.wrds.stop.n.log D.P.http
## 1965 2648
## D.P.mini D.P.air
## 2615 2627
## D.P.black D.P.white
## 2631 2638
## D.P.gold D.P.spacegray
## 2647 2642
## startprice.dgt1.is9 startprice.dgt2.is9
## 1779 2290
## startprice.dgt3.is9 startprice.dcm1.is9
## 2643 1653
## startprice.dcm2.is9
## 1826
## [1] "numeric data w/ Infs in : "
## named integer(0)
## [1] "numeric data w/ NaNs in : "
## named integer(0)
## [1] "string data missing in : "
## description condition cellular carrier color storage
## 1513 0 0 0 0 0
## productline .grpid descr.my
## 0 NA 1513# glb_allobs_df <- na.omit(glb_allobs_df)
# Not refactored into mydsutils.R since glb_*_df might be reassigned
glb_impute_missing_data <- function() {
require(mice)
set.seed(glb_mice_complete.seed)
inp_impent_df <- glb_allobs_df[, setdiff(names(glb_allobs_df),
union(glbFeatsExclude, glb_rsp_var))]
print("Summary before imputation: ")
print(summary(inp_impent_df))
out_impent_df <- complete(mice(inp_impent_df))
print(summary(out_impent_df))
ret_vars <- sapply(names(out_impent_df),
function(col) ifelse(!identical(out_impent_df[, col],
inp_impent_df[, col]),
col, ""))
ret_vars <- ret_vars[ret_vars != ""]
# complete(mice()) changes attributes of factors even though values don't change
for (col in ret_vars) {
if (inherits(out_impent_df[, col], "factor")) {
if (identical(as.numeric(out_impent_df[, col]),
as.numeric(inp_impent_df[, col])))
ret_vars <- setdiff(ret_vars, col)
}
}
return(out_impent_df[, ret_vars])
}
if (glb_impute_na_data &&
(length(myfind_numerics_missing(glb_allobs_df)) > 0) &&
(ncol(nonna_df <- glb_impute_missing_data()) > 0)) {
for (col in names(nonna_df)) {
glb_allobs_df[, paste0(col, ".nonNA")] <- nonna_df[, col]
glbFeatsExclude <- c(glbFeatsExclude, col)
}
}
mycheck_problem_data(glb_allobs_df, featsExclude = glbFeatsExclude,
fctrMaxUniqVals = glbFctrMaxUniqVals, terminate = TRUE)## [1] "numeric data missing in : "
## sold sold.fctr
## 798 798
## [1] "numeric data w/ 0s in : "
## biddable sold
## 1437 995
## sprice.log10 cellular.fctr
## 31 1590
## D.terms.post.stop.n D.terms.post.stop.n.log
## 1516 1516
## D.weight.post.stop.sum D.terms.post.stem.n
## 1516 1516
## D.terms.post.stem.n.log D.weight.post.stem.sum
## 1516 1516
## D.T.condit D.T.use
## 2156 2363
## D.T.in. D.T.good
## 2216 2451
## D.T.screen D.T.with
## 2440 2434
## D.T.of D.T.mint
## 2497 2585
## D.T.or D.T.cosmet
## 2523 2531
## D.T.minor D.T.light
## 2531 2567
## D.T.X100 D.T.from
## 2584 2591
## D.T.hous D.wrds.n.log
## 2576 1513
## D.wrds.unq.n.log D.weight.sum
## 1516 1516
## D.ratio.weight.sum.wrds.n D.chrs.n.log
## 1516 1513
## D.chrs.uppr.n.log D.dgts.n.log
## 1515 2452
## D.chrs.pnct01.n.log D.chrs.pnct02.n.log
## 2570 2648
## D.chrs.pnct03.n.log D.chrs.pnct04.n.log
## 2647 2648
## D.chrs.pnct05.n.log D.chrs.pnct06.n.log
## 2582 2596
## D.chrs.pnct07.n.log D.chrs.pnct08.n.log
## 2611 2572
## D.chrs.pnct09.n.log D.chrs.pnct10.n.log
## 2632 2639
## D.chrs.pnct11.n.log D.chrs.pnct12.n.log
## 2294 2534
## D.chrs.pnct13.n.log D.chrs.pnct14.n.log
## 1930 2574
## D.chrs.pnct15.n.log D.chrs.pnct16.n.log
## 2628 2639
## D.chrs.pnct17.n.log D.chrs.pnct18.n.log
## 2646 2647
## D.chrs.pnct19.n.log D.chrs.pnct20.n.log
## 2648 2648
## D.chrs.pnct21.n.log D.chrs.pnct22.n.log
## 2648 2648
## D.chrs.pnct23.n.log D.chrs.pnct24.n.log
## 2648 2648
## D.chrs.pnct25.n.log D.chrs.pnct26.n.log
## 2648 2648
## D.chrs.pnct27.n.log D.chrs.pnct28.n.log
## 2648 2640
## D.chrs.pnct29.n.log D.chrs.pnct30.n.log
## 2648 2648
## D.wrds.stop.n.log D.P.http
## 1965 2648
## D.P.mini D.P.air
## 2615 2627
## D.P.black D.P.white
## 2631 2638
## D.P.gold D.P.spacegray
## 2647 2642
## startprice.dgt1.is9 startprice.dgt2.is9
## 1779 2290
## startprice.dgt3.is9 startprice.dcm1.is9
## 2643 1653
## startprice.dcm2.is9
## 1826
## [1] "numeric data w/ Infs in : "
## named integer(0)
## [1] "numeric data w/ NaNs in : "
## named integer(0)
## [1] "string data missing in : "
## description condition cellular carrier color storage
## 1513 0 0 0 0 0
## productline .grpid descr.my
## 0 NA 1513Step 3.1: manage missing data
glb_chunks_df <- myadd_chunk(glb_chunks_df, "cluster.data", major.inc=FALSE)## label step_major step_minor label_minor bgn end
## 6 manage.missing.data 3 1 1 280.282 280.448
## 7 cluster.data 3 2 2 280.448 NA
## elapsed
## 6 0.166
## 7 NAmycompute_entropy_df <- function(obs_df, entropy_var, by_var=NULL) {
require(lazyeval)
require(dplyr)
require(tidyr)
if (is.null(by_var)) {
by_var <- ".default"
obs_df$.default <- as.factor(".default")
}
if (!any(grepl(".clusterid", names(obs_df), fixed=TRUE)))
obs_df$.clusterid <- 1
cluster_df <- obs_df %>%
count_(c(by_var, ".clusterid", entropy_var)) %>%
dplyr::filter(n > 0) %>%
dplyr::filter_(interp(~(!is.na(var)), var=as.name(entropy_var))) %>%
unite_(paste0(by_var, ".clusterid"),
c(interp(by_var), ".clusterid")) %>%
spread_(interp(entropy_var), "n", fill=0)
# head(cluster_df)
# sum(cluster_df$n)
tmp.entropy <- sapply(1:nrow(cluster_df),
function(row) entropy(as.numeric(cluster_df[row, -1]), method = "ML"))
tmp.knt <- sapply(1:nrow(cluster_df),
function(row) sum(as.numeric(cluster_df[row, -1])))
cluster_df$.entropy <- tmp.entropy; cluster_df$.knt <- tmp.knt
#print(cluster_df)
return(cluster_df)
}
if (glb_cluster) {
require(proxy)
#require(hash)
require(dynamicTreeCut)
require(entropy)
require(tidyr)
require(ggdendro)
mywgtdcosine_dist <- function(x, y=NULL, weights=NULL) {
if (!inherits(x, "matrix"))
x <- as.matrix(x)
if (is.null(weights))
weights <- rep(1, ncol(x))
wgtsx <- matrix(rep(weights / sum(weights), nrow(x)), nrow = nrow(x),
byrow = TRUE)
wgtdx <- x * wgtsx
wgtdxsqsum <- as.matrix(rowSums((x ^ 2) * wgtsx), byrow=FALSE)
denom <- sqrt(wgtdxsqsum %*% t(wgtdxsqsum))
ret_mtrx <- 1 - ((sum(weights) ^ 1) * (wgtdx %*% t(wgtdx)) / denom)
ret_mtrx[is.nan(ret_mtrx)] <- 1
diag(ret_mtrx) <- 0
return(ret_mtrx)
}
#pr_DB$delete_entry("mywgtdcosine");
# Need to do this only once across runs ?
if (!pr_DB$entry_exists("mywgtdcosine")) {
pr_DB$set_entry(FUN = mywgtdcosine_dist, names = c("mywgtdcosine"))
pr_DB$modify_entry(names="mywgtdcosine", type="metric", loop=FALSE)
}
#pr_DB$get_entry("mywgtdcosine")
# glb_hash <- hash(key=unique(glb_allobs_df$myCategory),
# values=1:length(unique(glb_allobs_df$myCategory)))
# glb_hash_lst <- hash(key=unique(glb_allobs_df$myCategory),
# values=1:length(unique(glb_allobs_df$myCategory)))
#stop(here"); glb_to_sav(); glb_allobs_df <- sav_allobs_df
cluster_vars <- grep(paste0("[",
toupper(paste0(substr(glbFeatsText, 1, 1), collapse = "")),
"]\\.[PT]\\."),
names(glb_allobs_df), value = TRUE)
# Assign correlations with rsp_var as weights for cosine distance
print("Clustering features: ")
cluster_vars_df <- data.frame(abs.cor.y = abs(cor(
glb_allobs_df[glb_allobs_df$.src == "Train", cluster_vars],
as.numeric(glb_allobs_df[glb_allobs_df$.src == "Train", glb_rsp_var]),
use = "pairwise.complete.obs")))
print(tail(cluster_vars_df <- orderBy(~ abs.cor.y,
subset(cluster_vars_df, !is.na(abs.cor.y))), 5))
print(sprintf(" .rnorm cor: %0.4f",
cor(glb_allobs_df[glb_allobs_df$.src == "Train", ".rnorm"],
as.numeric(glb_allobs_df[glb_allobs_df$.src == "Train", glb_rsp_var]),
use = "pairwise.complete.obs")))
print(sprintf("glb_allobs_df Entropy: %0.4f",
allobs_ent <- entropy(table(glb_allobs_df[, glb_cluster_entropy_var]),
method="ML")))
print(category_df <- mycompute_entropy_df(obs_df=glb_allobs_df,
entropy_var=glb_cluster_entropy_var,
by_var=glb_category_var))
print(sprintf("glb_allobs_df$%s Entropy: %0.4f (%0.4f pct)",
glb_category_var,
category_ent <- weighted.mean(category_df$.entropy, category_df$.knt),
100 * category_ent / allobs_ent))
glb_allobs_df$.clusterid <- 1
#print(max(table(glb_allobs_df$myCategory.fctr) / 20))
#stop(here"); glb_to_sav()
grp_ids <- sort(unique(glb_allobs_df[, glb_category_var]))
glb_cluster_size_df_lst <- list()
png(paste0(glb_out_pfx, "FeatsTxtClusters.png"),
width = 480 * 2, height = 480 * length(grp_ids))
grid.newpage()
pushViewport(viewport(layout = grid.layout(nrow = length(grp_ids), ncol = 2)))
pltIx <- 1
for (grp in grp_ids) {
# if (grep(grp, levels(grp_ids)) <= 6) next
# if (grep(grp, levels(grp_ids)) > 9) next
# if (grep(grp, levels(grp_ids)) != 10) next
print(sprintf("Category: %s", grp))
ctgry_allobs_df <- glb_allobs_df[glb_allobs_df[, glb_category_var] == grp, ]
if (!inherits(ctgry_allobs_df[, glb_cluster_entropy_var], "factor"))
ctgry_allobs_df[, glb_cluster_entropy_var] <-
as.factor(ctgry_allobs_df[, glb_cluster_entropy_var])
#dstns_dist <- proxy::dist(ctgry_allobs_df[, cluster_vars], method = "cosine")
dstns_dist <- proxy::dist(ctgry_allobs_df[, row.names(cluster_vars_df)],
method = "mywgtdcosine",
weights = cluster_vars_df$abs.cor.y)
# Custom distance functions return a crossdist object
#dstns_mtrx <- as.matrix(dstns_dist)
dstns_mtrx <- matrix(as.vector(dstns_dist), nrow=attr(dstns_dist, "dim")[1],
dimnames=attr(dstns_dist, "dimnames"))
dstns_dist <- as.dist(dstns_mtrx)
print(sprintf("max distance(%0.4f) pair:", max(dstns_mtrx)))
# print(dim(dstns_mtrx))
# print(sprintf("which.max: %d", which.max(dstns_mtrx)))
row_ix <- ceiling(which.max(dstns_mtrx) / ncol(dstns_mtrx))
col_ix <- which.max(dstns_mtrx[row_ix, ])
# print(sprintf("row_ix: %d", row_ix)); print(sprintf("col_ix: %d", col_ix));
# print(dim(ctgry_allobs_df))
print(ctgry_allobs_df[c(row_ix, col_ix),
c(glb_id_var, glb_cluster_entropy_var, glb_category_var, glbFeatsText, cluster_vars)])
min_dstns_mtrx <- dstns_mtrx
diag(min_dstns_mtrx) <- 1
# Float representations issue -2.22e-16 vs. 0.0000
print(sprintf("min distance(%0.4f) pair:", min(min_dstns_mtrx)))
row_ix <- ceiling(which.min(min_dstns_mtrx) / ncol(min_dstns_mtrx))
col_ix <- which.min(min_dstns_mtrx[row_ix, ])
print(ctgry_allobs_df[c(row_ix, col_ix),
c(glb_id_var, glb_cluster_entropy_var, glb_category_var, glbFeatsText,
cluster_vars)])
set.seed(glb_cluster.seed)
clusters <- hclust(dstns_dist, method = "ward.D2")
# Workaround to avoid "Error in cutree(dendro, h = heightcutoff) : the 'height' component of 'tree' is not sorted (increasingly)"
if (with(clusters,all.equal(height,sort(height))))
clusters$height <- round(clusters$height,6)
clusters$labels <- ctgry_allobs_df[, glb_id_var]
clustersDD <- dendro_data(clusters)
clustersDD$labels[, glb_rsp_var] <- sapply(clustersDD$labels$label, function(id)
ctgry_allobs_df[id == ctgry_allobs_df[, glb_id_var], glb_rsp_var])
print(ggdendrogram(clustersDD, rotate = TRUE, size = 2) +
geom_point(data = clustersDD$labels,
aes_string(x = "x", color = glb_rsp_var), y = min(clustersDD$segments$y)) +
coord_flip(ylim = c(min(clustersDD$segments$y),
max(clustersDD$segments$y))) +
ggtitle(grp),
vp = viewport(layout.pos.row = pltIx, layout.pos.col = 1))
# clusters$labels <- ctgry_allobs_df[, glb_id_var]
# clustersDD <- dendro_data(clusters)
# clustersDD$labels$color <- sapply(clustersDD$labels$label, function(id)
# ctgry_allobs_df[id == ctgry_allobs_df[, glb_id_var], glb_rsp_var])
# print(ggdendrogram(clustersDD, rotate = TRUE, size = 2) +
# geom_point(data = clustersDD$labels,
# aes_string(x = "x", color = "color"), y = min(clustersDD$segments$y)) +
# coord_flip(ylim = c(min(clustersDD$segments$y),
# max(clustersDD$segments$y))))
# print(ggdendrogram(clustersDD, rotate = TRUE, size = 2) +
# geom_point(data = clustersDD$labels,
# aes_string(x = "x", y = "y", color = "color")))
# myplclust(clusters, lab=ctgry_allobs_df[, glb_id_var],
# lab.col=unclass(ctgry_allobs_df[, glb_cluster_entropy_var]))
opt_minclustersize_df <- data.frame(minclustersize = nrow(ctgry_allobs_df),
entropy = entropy(table(ctgry_allobs_df[, glb_cluster_entropy_var]),
method = "ML"))
for (minclustersize in
as.integer(seq(nrow(ctgry_allobs_df) / 2, nrow(ctgry_allobs_df) / 10,
length = 5))) {
clusterGroups <- cutreeDynamic(clusters, minClusterSize = minclustersize,
method = "tree", deepSplit = 0)
# Unassigned groups are labeled 0; the largest group has label 1
clusterGroups[clusterGroups == 0] <- 1
ctgry_allobs_df$.clusterid <- clusterGroups
ctgry_clstrs_df <- mycompute_entropy_df(ctgry_allobs_df,
glb_cluster_entropy_var)
opt_minclustersize_df <- rbind(opt_minclustersize_df,
data.frame(minclustersize = minclustersize,
entropy = weighted.mean(ctgry_clstrs_df$.entropy, ctgry_clstrs_df$.knt)))
}
opt_minclustersize <-
opt_minclustersize_df$minclustersize[which.min(opt_minclustersize_df$entropy)]
opt_minclustersize_df$.color <-
ifelse(opt_minclustersize_df$minclustersize == opt_minclustersize,
"red", "blue")
print(ggplot(data = opt_minclustersize_df,
mapping = aes(x = minclustersize, y = entropy)) +
geom_point(aes(color = .color)) + scale_color_identity() +
guides(color = "none") + geom_line(),
vp = viewport(layout.pos.row = pltIx, layout.pos.col = 2))
glb_cluster_size_df_lst[[grp]] <- opt_minclustersize_df
# select minclustersize that minimizes entropy
clusterGroups <- cutreeDynamic(clusters, minClusterSize = opt_minclustersize,
method = "tree",
deepSplit = 0)
# Unassigned groups are labeled 0; the largest group has label 1
table(clusterGroups, ctgry_allobs_df[, glb_cluster_entropy_var],
useNA = "ifany")
clusterGroups[clusterGroups == 0] <- 1
table(clusterGroups, ctgry_allobs_df[, glb_cluster_entropy_var], useNA = "ifany")
glb_allobs_df[glb_allobs_df[, glb_category_var] == grp,]$.clusterid <-
clusterGroups
pltIx <- pltIx + 1
}
dev.off()
#all.equal(sav_allobs_df_clusterid, glb_allobs_df$.clusterid)
print(cluster_df <- mycompute_entropy_df(obs_df=glb_allobs_df,
entropy_var=glb_cluster_entropy_var,
by_var=glb_category_var))
print(sprintf("glb_allobs_df$%s$.clusterid Entropy: %0.4f (%0.4f pct)",
glb_category_var,
cluster_ent <- weighted.mean(cluster_df$.entropy, cluster_df$.knt),
100 * cluster_ent / category_ent))
glb_allobs_df$.clusterid.fctr <- as.factor(glb_allobs_df$.clusterid)
# .clusterid.fctr is created automatically (probably ?) later
glbFeatsExclude <- c(glbFeatsExclude, ".clusterid")
if (!is.null(glb_category_var))
# glb_interaction_only_feats_lst[ifelse(grepl("\\.fctr", glb_category_var),
# glb_category_var,
# paste0(glb_category_var, ".fctr"))] <-
# c(".clusterid.fctr")
glb_interaction_only_feats_lst[[".clusterid.fctr"]] <-
ifelse(grepl("\\.fctr", glb_category_var), glb_category_var,
paste0(glb_category_var, ".fctr"))
if (glbFeatsTextClusterVarsExclude)
glbFeatsExclude <- c(glbFeatsExclude, cluster_vars)
}## Loading required package: proxy
##
## Attaching package: 'proxy'
##
## The following objects are masked from 'package:stats':
##
## as.dist, dist
##
## The following object is masked from 'package:base':
##
## as.matrix
##
## Loading required package: dynamicTreeCut
## Loading required package: entropy
## Loading required package: ggdendro## [1] "Clustering features: "## Warning in cor(glb_allobs_df[glb_allobs_df$.src == "Train",
## cluster_vars], : the standard deviation is zero## abs.cor.y
## D.T.with 0.06608760
## D.T.in. 0.07165392
## D.T.cosmet 0.08848086
## D.T.X100 0.11139850
## D.T.hous 0.12927488
## [1] " .rnorm cor: -0.0013"
## [1] "glb_allobs_df Entropy: 0.6903"## Loading required package: lazyeval## Source: local data frame [10 x 5]
##
## prdl.my.fctr.clusterid N Y .entropy .knt
## (chr) (dbl) (dbl) (dbl) (dbl)
## 1 Unknown_1 122 82 0.6737987 204
## 2 iPad1_1 100 125 0.6869616 225
## 3 iPad2_1 139 147 0.6927559 286
## 4 iPad3_1 73 80 0.6921002 153
## 5 iPad4_1 93 64 0.6759893 157
## 6 iPadAir2_1 100 71 0.6786969 171
## 7 iPadAir_1 102 78 0.6842318 180
## 8 iPadmini2_1 58 49 0.6896056 107
## 9 iPadmini3_1 63 27 0.6108643 90
## 10 iPadmini_1 145 132 0.6920455 277
## [1] "glb_allobs_df$prdl.my.fctr Entropy: 0.6821 (98.8123 pct)"## [1] "Category: Unknown"
## [1] "max distance(1.0000) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 5 10005 N Unknown
## 24 10024 N Unknown
## descr.my
## 5 Please feel free to buy. All products have been thoroughly inspected, cleaned and tested to be 100%
## 24
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of D.T.mint
## 5 0 0 0 0 0 0 0 0
## 24 0 0 0 0 0 0 0 0
## D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from D.T.hous
## 5 0 0 0 0 0.4475573 0 0
## 24 0 0 0 0 0.0000000 0 0
## D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold D.P.spacegray
## 5 0 0 0 0 0 0 0
## 24 0 0 0 0 0 0 0
## [1] "min distance(0.8891) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 5 10005 N Unknown
## 419 10420 Y Unknown
## descr.my
## 5 Please feel free to buy. All products have been thoroughly inspected, cleaned and tested to be 100%
## 419 THIS ITEM WAS A STORE DEMO BUT HAS 100% BEEN RESTORED TO FACTORY SETTINGS AND IS READY FOR USE. DOES
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of
## 5 0 0.0000000 0 0 0 0 0
## 419 0 0.1891688 0 0 0 0 0
## D.T.mint D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from
## 5 0 0 0 0 0 0.4475573 0
## 419 0 0 0 0 0 0.3159228 0
## D.T.hous D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold
## 5 0 0 0 0 0 0 0
## 419 0 0 0 0 0 0 0
## D.P.spacegray
## 5 0
## 419 0## Warning in max(vapply(evaled, length, integer(1))): no non-missing
## arguments to max; returning -Inf## [1] "Category: iPad1"
## [1] "max distance(1.0000) pair:"
## UniqueID sold.fctr prdl.my.fctr descr.my D.T.condit D.T.use D.T.in.
## 9 10009 Y iPad1 0 0 0
## 12 10012 N iPad1 0 0 0
## D.T.good D.T.screen D.T.with D.T.of D.T.mint D.T.or D.T.cosmet
## 9 0 0 0 0 0 0 0
## 12 0 0 0 0 0 0 0
## D.T.minor D.T.light D.T.X100 D.T.from D.T.hous D.P.http D.P.mini
## 9 0 0 0 0 0 0 0
## 12 0 0 0 0 0 0 0
## D.P.air D.P.black D.P.white D.P.gold D.P.spacegray
## 9 0 0 0 0 0
## 12 0 0 0 0 0
## [1] "min distance(0.8976) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 2290 12301 <NA> iPad1
## 2326 12337 <NA> iPad1
## descr.my
## 2290 A device listed in Average/Fair condition with some scratches, scuffs on the housing & screen (do
## 2326 A device listed in Average/Fair condition with some scratches, scuffs on the housing & screen (do
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of
## 2290 0.1618782 0 0.1743867 0 0.2446832 0.241948 0
## 2326 0.1618782 0 0.1743867 0 0.2446832 0.241948 0
## D.T.mint D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from
## 2290 0 0 0 0 0 0 0
## 2326 0 0 0 0 0 0 0
## D.T.hous D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold
## 2290 0.3467175 0 0 0 0 0 0
## 2326 0.3467175 0 0 0 0 0 0
## D.P.spacegray
## 2290 0
## 2326 0## Warning in max(vapply(evaled, length, integer(1))): no non-missing
## arguments to max; returning -Inf## [1] "Category: iPad2"
## [1] "max distance(1.0000) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 1 10001 N iPad2
## 21 10021 Y iPad2
## descr.my D.T.condit D.T.use
## 1 iPad is in 8point5+ out of 10 cosmetic condition! 0.3035216 0
## 21 Crack on screen. 0.0000000 0
## D.T.in. D.T.good D.T.screen D.T.with D.T.of D.T.mint D.T.or
## 1 0.326975 0 0.000000 0 0.5165353 0 0
## 21 0.000000 0 1.223416 0 0.0000000 0 0
## D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from D.T.hous D.P.http
## 1 0.5625403 0 0 0 0 0 0
## 21 0.0000000 0 0 0 0 0 0
## D.P.mini D.P.air D.P.black D.P.white D.P.gold D.P.spacegray
## 1 0 0 0 0 0 0
## 21 0 0 0 0 0 0
## [1] "min distance(0.8886) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 431 10432 N iPad2
## 664 10665 N iPad2
## descr.my
## 431 *FREE* Same-Day Ship | Factory Refurbished | 90-Day Warranty | 100% Functional, Includes All Major
## 664 100% Fully Functional, Clean ESN & iCloud Clear, professionally restored, inspected, & tested. This
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of
## 431 0 0 0 0 0 0 0
## 664 0 0 0 0 0 0 0
## D.T.mint D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from
## 431 0 0 0 0 0 0.4882443 0
## 664 0 0 0 0 0 0.4882443 0
## D.T.hous D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold
## 431 0 0 0 0 0 0 0
## 664 0 0 0 0 0 0 0
## D.P.spacegray
## 431 0
## 664 0## Warning in max(vapply(evaled, length, integer(1))): no non-missing
## arguments to max; returning -Inf## [1] "No module detected"
## [1] "Category: iPad3"
## [1] "max distance(1.0000) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 17 10017 Y iPad3
## 37 10037 Y iPad3
## descr.my
## 17 Great working iPad. Very minor surface scratches on back as pictured. Other very light scratching
## 37 Rarely ever used it.
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of
## 17 0 0.000000 0 0 0 0 0
## 37 0 1.607935 0 0 0 0 0
## D.T.mint D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from
## 17 0 0 0 0.3214516 0.3593455 0 0
## 37 0 0 0 0.0000000 0.0000000 0 0
## D.T.hous D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold
## 17 0 0 0 0 0 0 0
## 37 0 0 0 0 0 0 0
## D.P.spacegray
## 17 0
## 37 0
## [1] "min distance(0.8886) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 946 10949 N iPad3
## 1910 11921 <NA> iPad3
## descr.my
## 946 *FREE* Same-Day Ship | 90-Day Warranty | 100% Functional, Includes All Major Accessories, Shows
## 1910 *FREE* Same-Day Ship | 90-Day Warranty | 100% Functional, Includes All Major Accessories, Shows
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of
## 946 0 0 0 0 0 0 0
## 1910 0 0 0 0 0 0 0
## D.T.mint D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from
## 946 0 0 0 0 0 0.4882443 0
## 1910 0 0 0 0 0 0.4882443 0
## D.T.hous D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold
## 946 0 0 0 0 0 0 0
## 1910 0 0 0 0 0 0 0
## D.P.spacegray
## 946 0
## 1910 0## Warning in max(vapply(evaled, length, integer(1))): no non-missing
## arguments to max; returning -Inf## [1] "Category: iPad4"
## [1] "max distance(1.0000) pair:"
## UniqueID sold.fctr prdl.my.fctr descr.my D.T.condit D.T.use D.T.in.
## 3 10003 Y iPad4 0 0 0
## 10 10010 Y iPad4 0 0 0
## D.T.good D.T.screen D.T.with D.T.of D.T.mint D.T.or D.T.cosmet
## 3 0 0 0 0 0 0 0
## 10 0 0 0 0 0 0 0
## D.T.minor D.T.light D.T.X100 D.T.from D.T.hous D.P.http D.P.mini
## 3 0 0 0 0 0 0 0
## 10 0 0 0 0 0 0 0
## D.P.air D.P.black D.P.white D.P.gold D.P.spacegray
## 3 0 0 0 0 0
## 10 0 0 0 0 0
## [1] "min distance(0.8886) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 63 10063 N iPad4
## 2308 12319 <NA> iPad4
## descr.my
## 63 *FREE* Same-Day Ship | 90-Day Warranty | 100% Functional, Includes All Major Accessories, Shows
## 2308 *FREE* Same-Day Ship | 90-Day Warranty | 100% Functional, Includes All Major Accessories, Shows
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of
## 63 0 0 0 0 0 0 0
## 2308 0 0 0 0 0 0 0
## D.T.mint D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from
## 63 0 0 0 0 0 0.4882443 0
## 2308 0 0 0 0 0 0.4882443 0
## D.T.hous D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold
## 63 0 0 0 0 0 0 0
## 2308 0 0 0 0 0 0 0
## D.P.spacegray
## 63 0
## 2308 0## Warning in max(vapply(evaled, length, integer(1))): no non-missing
## arguments to max; returning -Inf## [1] "Category: iPadAir"
## [1] "max distance(1.0000) pair:"
## UniqueID sold.fctr prdl.my.fctr descr.my D.T.condit D.T.use D.T.in.
## 16 10016 N iPadAir 0 0 0
## 25 10025 N iPadAir 0 0 0
## D.T.good D.T.screen D.T.with D.T.of D.T.mint D.T.or D.T.cosmet
## 16 0 0 0 0 0 0 0
## 25 0 0 0 0 0 0 0
## D.T.minor D.T.light D.T.X100 D.T.from D.T.hous D.P.http D.P.mini
## 16 0 0 0 0 0 0 0
## 25 0 0 0 0 0 0 0
## D.P.air D.P.black D.P.white D.P.gold D.P.spacegray
## 16 0 0 0 0 0
## 25 0 0 0 0 0
## [1] "min distance(0.8886) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 167 10167 N iPadAir
## 1156 11162 N iPadAir
## descr.my
## 167 LIKE BRANDNEW 100% MONEY BACK GUARANTEE
## 1156 *FREE* Same-Day Ship | 90-Day Warranty | 100% Functional, Includes All Major Accessories, Shows
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of
## 167 0 0 0 0 0 0 0
## 1156 0 0 0 0 0 0 0
## D.T.mint D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from
## 167 0 0 0 0 0 1.7902291 0
## 1156 0 0 0 0 0 0.4882443 0
## D.T.hous D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold
## 167 0 0 0 0 0 0 0
## 1156 0 0 0 0 0 0 0
## D.P.spacegray
## 167 0
## 1156 0## Warning in max(vapply(evaled, length, integer(1))): no non-missing
## arguments to max; returning -Inf## [1] "Category: iPadAir2"
## [1] "max distance(1.0000) pair:"
## UniqueID sold.fctr prdl.my.fctr descr.my D.T.condit D.T.use D.T.in.
## 19 10019 Y iPadAir2 0 0 0
## 35 10035 N iPadAir2 0 0 0
## D.T.good D.T.screen D.T.with D.T.of D.T.mint D.T.or D.T.cosmet
## 19 0 0 0 0 0 0 0
## 35 0 0 0 0 0 0 0
## D.T.minor D.T.light D.T.X100 D.T.from D.T.hous D.P.http D.P.mini
## 19 0 0 0 0 0 0 0
## 35 0 0 0 0 0 0 0
## D.P.air D.P.black D.P.white D.P.gold D.P.spacegray
## 19 0 0 0 0 0
## 35 0 0 0 0 0
## [1] "min distance(0.9283) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 1214 11220 N iPadAir2
## 1856 11867 <NA> iPadAir2
## descr.my
## 1214 Brandnew, iPad untouched and still wrapped in original plastic inside box. Opened only to verify
## 1856 Brandnew factory sealed iPad in an OPEN BOX...THE BOX ITSELF IS HEAVILY DISTRESSED(see
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of
## 1214 0 0 0.2378000 0 0 0 0
## 1856 0 0 0.2012154 0 0 0 0
## D.T.mint D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from
## 1214 0 0 0 0 0 0 0
## 1856 0 0 0 0 0 0 0
## D.T.hous D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold
## 1214 0 0 0 0 0 0 0
## 1856 0 0 0 0 0 0 0
## D.P.spacegray
## 1214 0
## 1856 0## Warning in max(vapply(evaled, length, integer(1))): no non-missing
## arguments to max; returning -Inf## [1] "No module detected"
## [1] "Category: iPadmini"
## [1] "max distance(1.0000) pair:"
## UniqueID sold.fctr prdl.my.fctr descr.my D.T.condit D.T.use D.T.in.
## 7 10007 Y iPadmini 0 0 0
## 57 10057 N iPadmini 0 0 0
## D.T.good D.T.screen D.T.with D.T.of D.T.mint D.T.or D.T.cosmet
## 7 0 0 0 0 0 0 0
## 57 0 0 0 0 0 0 0
## D.T.minor D.T.light D.T.X100 D.T.from D.T.hous D.P.http D.P.mini
## 7 0 0 0 0 0 0 0
## 57 0 0 0 0 0 0 0
## D.P.air D.P.black D.P.white D.P.gold D.P.spacegray
## 7 0 0 0 0 0
## 57 0 0 0 0 0
## [1] "min distance(0.8919) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 93 10093 N iPadmini
## 2364 12375 <NA> iPadmini
## descr.my
## 93 *FREE* Same-Day Ship | 90-Day Warranty | 100% Functional, Includes All Major Accessories, Shows
## 2364 100% working condition, No icloud blocks, tablet is ready for new user. See item description for
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of
## 93 0.0000000 0 0 0 0 0 0
## 2364 0.1734409 0 0 0 0 0 0
## D.T.mint D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from
## 93 0 0 0 0 0 0.4882443 0
## 2364 0 0 0 0 0 0.3836205 0
## D.T.hous D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold
## 93 0 0 0 0 0 0 0
## 2364 0 0 0 0 0 0 0
## D.P.spacegray
## 93 0
## 2364 0## Warning in max(vapply(evaled, length, integer(1))): no non-missing
## arguments to max; returning -Inf## [1] "No module detected"
## [1] "Category: iPadmini2"
## [1] "max distance(1.0000) pair:"
## UniqueID sold.fctr prdl.my.fctr descr.my D.T.condit D.T.use D.T.in.
## 4 10004 N iPadmini2 0 0 0
## 6 10006 Y iPadmini2 0 0 0
## D.T.good D.T.screen D.T.with D.T.of D.T.mint D.T.or D.T.cosmet D.T.minor
## 4 0 0 0 0 0 0 0 0
## 6 0 0 0 0 0 0 0 0
## D.T.light D.T.X100 D.T.from D.T.hous D.P.http D.P.mini D.P.air D.P.black
## 4 0 0 0 0 0 0 0 0
## 6 0 0 0 0 0 0 0 0
## D.P.white D.P.gold D.P.spacegray
## 4 0 0 0
## 6 0 0 0
## [1] "min distance(0.9126) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 927 10930 N iPadmini2
## 2020 12031 <NA> iPadmini2
## descr.my
## 927 Flawless Condition(10/10) 100% functional with Flawless Retina Display. Unit is free of
## 2020 Good Condition(8point25/10), 100% functional with Flawless Retina Display. Unit has a dent on upper left
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of
## 927 0.1867825 0 0 0.0000000 0 0.2791708 0.3178679
## 2020 0.1734409 0 0 0.2677597 0 0.2592300 0.0000000
## D.T.mint D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from
## 927 0 0 0 0 0 0.4131298 0
## 2020 0 0 0 0 0 0.3836205 0
## D.T.hous D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold
## 927 0 0 0 0 0 0 0
## 2020 0 0 0 0 0 0 0
## D.P.spacegray
## 927 0
## 2020 0## Warning in max(vapply(evaled, length, integer(1))): no non-missing
## arguments to max; returning -Inf## [1] "Category: iPadmini3"
## [1] "max distance(1.0000) pair:"
## UniqueID sold.fctr prdl.my.fctr descr.my D.T.condit D.T.use D.T.in.
## 8 10008 N iPadmini3 0 0 0
## 120 10120 N iPadmini3 0 0 0
## D.T.good D.T.screen D.T.with D.T.of D.T.mint D.T.or D.T.cosmet
## 8 0 0 0 0 0 0 0
## 120 0 0 0 0 0 0 0
## D.T.minor D.T.light D.T.X100 D.T.from D.T.hous D.P.http D.P.mini
## 8 0 0 0 0 0 0 0
## 120 0 0 0 0 0 0 0
## D.P.air D.P.black D.P.white D.P.gold D.P.spacegray
## 8 0 0 0 0 0
## 120 0 0 0 0 0
## [1] "min distance(0.9488) pair:"
## UniqueID sold.fctr prdl.my.fctr
## 1011 11014 N iPadmini3
## 1115 11121 N iPadmini3
## descr.my
## 1011 Brandnew IpadMini, still in plastic wrap and box provided by hospital with cords and new unopened
## 1115 Bought for one day and removed iPad plastic cover, never used. In original box with all accessories!
## D.T.condit D.T.use D.T.in. D.T.good D.T.screen D.T.with D.T.of
## 1011 0 0.0000000 0.2179833 0 0 0.302435 0
## 1115 0 0.2143913 0.1743867 0 0 0.241948 0
## D.T.mint D.T.or D.T.cosmet D.T.minor D.T.light D.T.X100 D.T.from
## 1011 0 0 0 0 0 0 0
## 1115 0 0 0 0 0 0 0
## D.T.hous D.P.http D.P.mini D.P.air D.P.black D.P.white D.P.gold
## 1011 0 0 1 0 0 0 0
## 1115 0 0 0 0 0 0 0
## D.P.spacegray
## 1011 0
## 1115 0## Warning in max(vapply(evaled, length, integer(1))): no non-missing
## arguments to max; returning -Inf## [1] "No module detected"
## Source: local data frame [19 x 5]
##
## prdl.my.fctr.clusterid N Y .entropy .knt
## (chr) (dbl) (dbl) (dbl) (dbl)
## 1 Unknown_1 92 62 0.6740508 154
## 2 Unknown_2 17 12 0.6782094 29
## 3 Unknown_3 13 8 0.6645284 21
## 4 iPad1_1 72 104 0.6765260 176
## 5 iPad1_2 15 12 0.6869616 27
## 6 iPad1_3 13 9 0.6765260 22
## 7 iPad2_1 139 147 0.6927559 286
## 8 iPad3_1 60 75 0.6869616 135
## 9 iPad3_2 13 5 0.5908422 18
## 10 iPad4_1 52 52 0.6931472 104
## 11 iPad4_2 16 12 0.6829081 28
## 12 iPad4_3 25 0 0.0000000 25
## 13 iPadAir2_1 100 71 0.6786969 171
## 14 iPadAir_1 86 60 0.6772057 146
## 15 iPadAir_2 16 18 0.6914161 34
## 16 iPadmini2_1 50 44 0.6911087 94
## 17 iPadmini2_2 8 5 0.6662784 13
## 18 iPadmini3_1 63 27 0.6108643 90
## 19 iPadmini_1 145 132 0.6920455 277
## [1] "glb_allobs_df$prdl.my.fctr$.clusterid Entropy: 0.6710 (98.3773 pct)"# Last call for data modifications
#stop(here") # sav_allobs_df <- glb_allobs_df
# glb_allobs_df[(glb_allobs_df$PropR == 0.75) & (glb_allobs_df$State == "Hawaii"), "PropR.fctr"] <- "N"
# Re-partition
glb_trnobs_df <- subset(glb_allobs_df, .src == "Train")
glb_newobs_df <- subset(glb_allobs_df, .src == "Test")
glb_chunks_df <- myadd_chunk(glb_chunks_df, "partition.data.training", major.inc=TRUE)## label step_major step_minor label_minor bgn
## 7 cluster.data 3 2 2 280.448
## 8 partition.data.training 4 0 0 330.238
## end elapsed
## 7 330.237 49.789
## 8 NA NAStep 4.0: partition data training
#stop(here"); glb_to_sav()
if (all(is.na(glb_newobs_df[, glb_rsp_var]))) {
set.seed(glb_split_sample.seed)
OOB_size <- nrow(glb_newobs_df) * 1.1
if (is.null(glb_category_var)) {
require(caTools)
split <- sample.split(glb_trnobs_df[, glb_rsp_var_raw],
SplitRatio=OOB_size / nrow(glb_trnobs_df))
glb_OOBobs_df <- glb_trnobs_df[split ,]
glb_fitobs_df <- glb_trnobs_df[!split, ]
} else {
sample_vars <- c(glb_category_var, glb_rsp_var_raw)
rspvar_freq_df <- orderBy(reformulate(glb_rsp_var_raw),
mycreate_sqlxtab_df(glb_trnobs_df, glb_rsp_var_raw))
OOB_rspvar_size <-
1.0 * OOB_size * rspvar_freq_df$.n / sum(rspvar_freq_df$.n)
names(OOB_rspvar_size) <- as.character(rspvar_freq_df[, glb_rsp_var_raw])
newobs_freq_df <- orderBy(reformulate(glb_category_var),
mycreate_sqlxtab_df(glb_newobs_df, glb_category_var))
trnobs_freq_df <- orderBy(reformulate(glb_category_var),
mycreate_sqlxtab_df(glb_trnobs_df, glb_category_var))
ctgry_freq_df <- merge(newobs_freq_df, trnobs_freq_df,
by = glb_category_var,
all = TRUE, sort = TRUE,
suffixes = c(".tst", ".trn"))
ctgry_freq_df[is.na(ctgry_freq_df)] <- 0
obs_freq_df <- mycreate_xtab_df(glb_trnobs_df,
c(glb_category_var, glb_rsp_var_raw))
newobs_freq_df <- orderBy(reformulate(glb_category_var),
mycreate_sqlxtab_df(glb_newobs_df, glb_category_var))
names(newobs_freq_df) <- gsub(".n", ".n.tst", names(newobs_freq_df),
fixed = TRUE)
obs_freq_df <- merge(obs_freq_df, newobs_freq_df, all = TRUE)
strata_mtrx <- ceiling(
matrix(obs_freq_df$.n.tst * 1.0 / sum(obs_freq_df$.n.tst)) %*%
matrix(OOB_rspvar_size, nrow = 1))
dimnames(strata_mtrx)[[1]] <- obs_freq_df[, glb_category_var]
dimnames(strata_mtrx)[[2]] <- as.character(rspvar_freq_df[, glb_rsp_var_raw])
for (val in rspvar_freq_df[, glb_rsp_var_raw]) {
trn <- paste0(glb_rsp_var_raw, ".", as.character(val))
strata <- paste0(".strata.", as.character(val))
obs_freq_df[, strata] <- strata_mtrx[, as.character(val)]
if (length((ix <- which(obs_freq_df[, trn] <
2.0 * obs_freq_df[, strata]))) > 0)
# More obs in OOB compared to fit currently
obs_freq_df[ix, strata] <- floor(obs_freq_df[ix, trn] / 2.0)
if (length((ix <- which(obs_freq_df[, strata] == 0))) > 0)
obs_freq_df[ix, strata] <- 1
}
#print(colSums(obs_freq_df[, -1]))
# obs_freq_df <- expand.grid(
# glb_category_var = ctgry_freq_df[, glb_category_var],
# glb_rsp_var_raw = rspvar_freq_df[, glb_rsp_var_raw])
# names(obs_freq_df) <- sample_vars
# trnobs_freq_df <- orderBy(reformulate(sample_vars),
# mycreate_sqlxtab_df(glb_trnobs_df, sample_vars))
# names(trnobs_freq_df) <- gsub(".n", ".n.trn", names(trnobs_freq_df),
# fixed = TRUE)
# obs_freq_df <- merge(obs_freq_df, trnobs_freq_df, all = TRUE)
# newobs_freq_df <- orderBy(reformulate(glb_category_var),
# mycreate_sqlxtab_df(glb_newobs_df, glb_category_var))
# # Since glb_rsp_var_raw might be NA in newobs_df, repeat .n.Test across unique values of glb_rsp_var_raw to prep obs_freq_df
# newobs_rsp_freq_df <- data.frame()
# for (val in rspvar_freq_df[, glb_rsp_var_raw])
# newobs_rsp_freq_df <- rbind(newobs_rsp_freq_df,
# cbind(newobs_freq_df, data.frame(glb_rsp_var_raw = val)))
# obs_freq_df <- merge(obs_freq_df, newobs_freq_df, all = TRUE)
# obs_freq_df <- orderBy(reformulate(sample_vars), obs_freq_df)
#
# obs_freq_df$.n.strata <- ceiling(
# as.vector(matrix(ctgry_freq_df$.n.Tst * 1.0 /
# sum(ctgry_freq_df$.n.Tst)) %*%
# matrix(OOB_rspvar_size, nrow = 1)))
# obs_freq_df[obs_freq_df$.n.Strata == 0, ".n.Strata"] <- 1
#
# # Adjust OOB_strata_size (desired # of OOB obs) if > # of trn obs
# ix <- which(!is.na(trnobs_univ_df$.n) &
# (OOB_strata_size > trnobs_univ_df$.n))
# if (length(ix) > 0)
# OOB_strata_size[ix] <- trnobs_univ_df[ix, ".n"]
# # Adjust OOB_strata_size (desired # of OOB obs) if > # of (trn - OOB) obs
# ix <- which(!is.na(trnobs_univ_df$.n) &
# (OOB_strata_size * 2 > trnobs_univ_df$.n))
# if (length(ix) > 0)
# OOB_strata_size[ix] <- trnobs_univ_df[ix, ".n"]
OOB_strata_size <- as.vector(as.matrix(obs_freq_df[,
grepl("^\\.strata\\.", names(obs_freq_df))]))
tmp_trnobs_df <- orderBy(reformulate(c(glb_rsp_var_raw, glb_category_var)),
glb_trnobs_df)
require(sampling)
split_strata <- sampling::strata(tmp_trnobs_df,
stratanames = c(glb_rsp_var_raw, glb_category_var),
size = OOB_strata_size[!is.na(OOB_strata_size)],
method = "srswor")
glb_OOBobs_df <- getdata(tmp_trnobs_df, split_strata)[, names(glb_trnobs_df)]
glb_fitobs_df <- glb_trnobs_df[!glb_trnobs_df[, glb_id_var] %in%
glb_OOBobs_df[, glb_id_var], ]
}
} else {
print(sprintf("Newdata contains non-NA data for %s; setting OOB to Newdata",
glb_rsp_var))
glb_fitobs_df <- glb_trnobs_df; glb_OOBobs_df <- glb_newobs_df
}## Loading required package: sampling
##
## Attaching package: 'sampling'
##
## The following objects are masked from 'package:survival':
##
## cluster, strata
##
## The following object is masked from 'package:caret':
##
## clusterif (!is.null(glb_max_fitobs) && (nrow(glb_fitobs_df) > glb_max_fitobs)) {
warning("glb_fitobs_df restricted to glb_max_fitobs: ",
format(glb_max_fitobs, big.mark = ","))
org_fitobs_df <- glb_fitobs_df
glb_fitobs_df <-
org_fitobs_df[split <- sample.split(org_fitobs_df[, glb_rsp_var_raw],
SplitRatio = glb_max_fitobs), ]
org_fitobs_df <- NULL
}
glb_allobs_df$.lcn <- ""; glb_trnobs_df$.lcn <- "";
glb_allobs_df[glb_allobs_df[, glb_id_var] %in%
glb_fitobs_df[, glb_id_var], ".lcn"] <- "Fit"
glb_trnobs_df[glb_trnobs_df[, glb_id_var] %in%
glb_fitobs_df[, glb_id_var], ".lcn"] <- "Fit"
glb_allobs_df[glb_allobs_df[, glb_id_var] %in%
glb_OOBobs_df[, glb_id_var], ".lcn"] <- "OOB"
glb_trnobs_df[glb_trnobs_df[, glb_id_var] %in%
glb_OOBobs_df[, glb_id_var], ".lcn"] <- "OOB"
dsp_class_dstrb <- function(obs_df, location_var, partition_var) {
xtab_df <- mycreate_xtab_df(obs_df, c(location_var, partition_var))
rownames(xtab_df) <- xtab_df[, location_var]
xtab_df <- xtab_df[, -grepl(location_var, names(xtab_df))]
print(xtab_df)
print(xtab_df / rowSums(xtab_df, na.rm=TRUE))
}
# Ensure proper splits by glb_rsp_var_raw & user-specified feature for OOB vs. new
if (!is.null(glb_category_var)) {
if (glb_is_classification)
dsp_class_dstrb(glb_allobs_df, ".lcn", glb_rsp_var_raw)
newobs_ctgry_df <- mycreate_sqlxtab_df(subset(glb_allobs_df, .src == "Test"),
glb_category_var)
names(newobs_ctgry_df) <-
gsub(".n", ".n.Tst", names(newobs_ctgry_df), fixed = TRUE)
OOBobs_ctgry_df <- mycreate_sqlxtab_df(subset(glb_allobs_df, .lcn == "OOB"),
glb_category_var)
names(OOBobs_ctgry_df) <-
gsub(".n", ".n.OOB", names(OOBobs_ctgry_df), fixed = TRUE)
fitobs_ctgry_df <- mycreate_sqlxtab_df(subset(glb_allobs_df, .lcn == "Fit"),
glb_category_var)
names(fitobs_ctgry_df) <-
gsub(".n", ".n.Fit", names(fitobs_ctgry_df), fixed = TRUE)
# glb_ctgry_df <- merge(newobs_ctgry_df, OOBobs_ctgry_df, by=glb_category_var
# , all=TRUE, suffixes=c(".Tst", ".OOB"))
glb_ctgry_df <- merge(fitobs_ctgry_df, OOBobs_ctgry_df, by = glb_category_var
, all = TRUE)
glb_ctgry_df <- merge(glb_ctgry_df, newobs_ctgry_df, by = glb_category_var
, all = TRUE)
glb_ctgry_df$.freqRatio.Fit <-
glb_ctgry_df$.n.Fit / sum(glb_ctgry_df$.n.Fit, na.rm = TRUE)
glb_ctgry_df$.freqRatio.OOB <-
glb_ctgry_df$.n.OOB / sum(glb_ctgry_df$.n.OOB, na.rm = TRUE)
glb_ctgry_df$.freqRatio.Tst <-
glb_ctgry_df$.n.Tst / sum(glb_ctgry_df$.n.Tst, na.rm = TRUE)
print(orderBy(~-.freqRatio.Tst-.freqRatio.OOB-.freqRatio.Fit, glb_ctgry_df))
}## sold.0 sold.1 sold.NA
## NA NA 798
## Fit 548 472 NA
## OOB 447 383 NA
## sold.0 sold.1 sold.NA
## NA NA 1
## Fit 0.5372549 0.4627451 NA
## OOB 0.5385542 0.4614458 NA
## prdl.my.fctr .n.Fit .n.OOB .n.Tst .freqRatio.Fit .freqRatio.OOB
## 3 iPad2 144 142 154 0.14117647 0.17108434
## 8 iPadmini 154 123 111 0.15098039 0.14819277
## 1 Unknown 108 96 92 0.10588235 0.11566265
## 2 iPad1 130 95 88 0.12745098 0.11445783
## 6 iPadAir 98 82 74 0.09607843 0.09879518
## 5 iPad4 84 73 68 0.08235294 0.08795181
## 7 iPadAir2 102 69 62 0.10000000 0.08313253
## 9 iPadmini2 54 53 56 0.05294118 0.06385542
## 4 iPad3 92 61 55 0.09019608 0.07349398
## 10 iPadmini3 54 36 38 0.05294118 0.04337349
## .freqRatio.Tst
## 3 0.19298246
## 8 0.13909774
## 1 0.11528822
## 2 0.11027569
## 6 0.09273183
## 5 0.08521303
## 7 0.07769424
## 9 0.07017544
## 4 0.06892231
## 10 0.04761905print("glb_allobs_df: "); print(dim(glb_allobs_df))## [1] "glb_allobs_df: "## [1] 2648 105print("glb_trnobs_df: "); print(dim(glb_trnobs_df))## [1] "glb_trnobs_df: "## [1] 1850 105print("glb_fitobs_df: "); print(dim(glb_fitobs_df))## [1] "glb_fitobs_df: "## [1] 1020 104print("glb_OOBobs_df: "); print(dim(glb_OOBobs_df))## [1] "glb_OOBobs_df: "## [1] 830 104print("glb_newobs_df: "); print(dim(glb_newobs_df))## [1] "glb_newobs_df: "## [1] 798 104# # Does not handle NULL or length(glb_id_var) > 1
if (glb_save_envir)
save(glb_allobs_df, #glb_trnobs_df, glb_fitobs_df, glb_OOBobs_df, glb_newobs_df,
file=paste0(glb_out_pfx, "blddfs_dsk.RData"))
# load(paste0(glb_out_pfx, "blddfs_dsk.RData"))
rm(split)## Warning in rm(split): object 'split' not foundglb_chunks_df <- myadd_chunk(glb_chunks_df, "select.features", major.inc=TRUE)## label step_major step_minor label_minor bgn
## 8 partition.data.training 4 0 0 330.238
## 9 select.features 5 0 0 331.640
## end elapsed
## 8 331.639 1.401
## 9 NA NAStep 5.0: select features
#stop(here"); glb_to_sav(); glb_allobs_df <- sav_allobs_df
print(glb_feats_df <- myselect_features(entity_df=glb_trnobs_df,
exclude_vars_as_features=glbFeatsExclude,
rsp_var=glb_rsp_var))## Warning in cor(data.matrix(entity_df[, sel_feats]), y =
## as.numeric(entity_df[, : the standard deviation is zero## id cor.y
## sold sold 1.000000000
## biddable biddable 0.549242842
## sprice.root2 sprice.root2 -0.511275385
## sprice.log10 sprice.log10 -0.469398937
## startprice startprice -0.458981936
## startprice.log10.predict startprice.log10.predict -0.408722051
## sprice.d20nexp sprice.d20nexp 0.397995133
## spdiff.cut.fctr spdiff.cut.fctr 0.290604966
## sprice.predict.diff sprice.predict.diff 0.271458109
## UniqueID UniqueID -0.190626150
## condition.fctr condition.fctr -0.155283820
## startprice.dgt1.is9 startprice.dgt1.is9 -0.138476687
## D.T.hous D.T.hous -0.129274878
## D.chrs.pnct05.n.log D.chrs.pnct05.n.log -0.120173065
## D.T.X100 D.T.X100 -0.111398502
## D.chrs.pnct06.n.log D.chrs.pnct06.n.log -0.095888068
## D.T.cosmet D.T.cosmet -0.088480863
## D.dgts.n.log D.dgts.n.log -0.082741631
## .clusterid .clusterid -0.080966421
## .clusterid.fctr .clusterid.fctr -0.080966421
## D.chrs.pnct14.n.log D.chrs.pnct14.n.log -0.075911474
## D.terms.post.stop.n D.terms.post.stop.n -0.075731613
## D.terms.post.stem.n D.terms.post.stem.n -0.074990216
## cellular.fctr cellular.fctr -0.073025939
## D.T.in. D.T.in. -0.071653919
## D.T.with D.T.with -0.066087601
## D.terms.post.stop.n.log D.terms.post.stop.n.log -0.062526435
## D.terms.post.stem.n.log D.terms.post.stem.n.log -0.062220631
## D.wrds.unq.n.log D.wrds.unq.n.log -0.062220631
## D.chrs.pnct09.n.log D.chrs.pnct09.n.log -0.061919863
## D.ratio.wrds.stop.n.wrds.n D.ratio.wrds.stop.n.wrds.n 0.059547328
## carrier.fctr carrier.fctr -0.058701911
## D.T.mint D.T.mint -0.055691207
## D.wrds.n.log D.wrds.n.log -0.055678363
## D.chrs.n.log D.chrs.n.log -0.055576665
## D.chrs.uppr.n.log D.chrs.uppr.n.log -0.054491610
## prdl.my.fctr prdl.my.fctr -0.054151937
## D.chrs.pnct28.n.log D.chrs.pnct28.n.log -0.052538378
## D.T.of D.T.of 0.052259530
## D.chrs.pnct12.n.log D.chrs.pnct12.n.log -0.049684867
## D.chrs.pnct15.n.log D.chrs.pnct15.n.log 0.048611857
## startprice.dcm2.is9 startprice.dcm2.is9 0.047716416
## D.weight.post.stem.sum D.weight.post.stem.sum -0.047105442
## D.weight.sum D.weight.sum -0.047105442
## D.weight.post.stop.sum D.weight.post.stop.sum -0.045009208
## color.fctr color.fctr -0.044231522
## D.terms.n.stem.stop.Ratio D.terms.n.stem.stop.Ratio 0.043562113
## startprice.dgt3.is9 startprice.dgt3.is9 -0.043150483
## D.T.minor D.T.minor -0.042692360
## D.T.screen D.T.screen 0.040605274
## startprice.dgt2.is9 startprice.dgt2.is9 -0.039616843
## D.chrs.pnct08.n.log D.chrs.pnct08.n.log -0.039606364
## D.T.condit D.T.condit -0.033248924
## D.T.light D.T.light -0.032814746
## D.chrs.pnct13.n.log D.chrs.pnct13.n.log -0.032095777
## D.T.from D.T.from -0.030653359
## D.chrs.pnct17.n.log D.chrs.pnct17.n.log 0.025087653
## D.chrs.pnct07.n.log D.chrs.pnct07.n.log 0.024330390
## D.chrs.pnct10.n.log D.chrs.pnct10.n.log -0.024107007
## D.T.or D.T.or -0.022614420
## D.chrs.pnct03.n.log D.chrs.pnct03.n.log -0.021557732
## D.chrs.pnct18.n.log D.chrs.pnct18.n.log -0.021557732
## D.P.gold D.P.gold -0.021557732
## D.chrs.pnct11.n.log D.chrs.pnct11.n.log -0.020186238
## D.P.white D.P.white 0.018597281
## D.T.good D.T.good 0.016394675
## D.chrs.pnct16.n.log D.chrs.pnct16.n.log 0.015642041
## storage.fctr storage.fctr -0.014549906
## D.P.air D.P.air -0.009203749
## startprice.dcm1.is9 startprice.dcm1.is9 -0.008364097
## D.P.mini D.P.mini -0.006894908
## D.chrs.pnct01.n.log D.chrs.pnct01.n.log 0.004291712
## D.wrds.stop.n.log D.wrds.stop.n.log 0.003851183
## D.P.spacegray D.P.spacegray 0.003532788
## D.T.use D.T.use 0.003025811
## D.weight.sum.stem.stop.Ratio D.weight.sum.stem.stop.Ratio 0.002381527
## .rnorm .rnorm -0.001273806
## D.P.black D.P.black -0.001181544
## D.ratio.weight.sum.wrds.n D.ratio.weight.sum.wrds.n -0.001080251
## D.chrs.pnct02.n.log D.chrs.pnct02.n.log NA
## D.chrs.pnct04.n.log D.chrs.pnct04.n.log NA
## D.chrs.pnct19.n.log D.chrs.pnct19.n.log NA
## D.chrs.pnct20.n.log D.chrs.pnct20.n.log NA
## D.chrs.pnct21.n.log D.chrs.pnct21.n.log NA
## D.chrs.pnct22.n.log D.chrs.pnct22.n.log NA
## D.chrs.pnct23.n.log D.chrs.pnct23.n.log NA
## D.chrs.pnct24.n.log D.chrs.pnct24.n.log NA
## D.chrs.pnct25.n.log D.chrs.pnct25.n.log NA
## D.chrs.pnct26.n.log D.chrs.pnct26.n.log NA
## D.chrs.pnct27.n.log D.chrs.pnct27.n.log NA
## D.chrs.pnct29.n.log D.chrs.pnct29.n.log NA
## D.chrs.pnct30.n.log D.chrs.pnct30.n.log NA
## D.P.http D.P.http NA
## exclude.as.feat cor.y.abs
## sold 1 1.000000000
## biddable 0 0.549242842
## sprice.root2 0 0.511275385
## sprice.log10 0 0.469398937
## startprice 1 0.458981936
## startprice.log10.predict 1 0.408722051
## sprice.d20nexp 0 0.397995133
## spdiff.cut.fctr 0 0.290604966
## sprice.predict.diff 1 0.271458109
## UniqueID 1 0.190626150
## condition.fctr 0 0.155283820
## startprice.dgt1.is9 0 0.138476687
## D.T.hous 0 0.129274878
## D.chrs.pnct05.n.log 0 0.120173065
## D.T.X100 0 0.111398502
## D.chrs.pnct06.n.log 0 0.095888068
## D.T.cosmet 0 0.088480863
## D.dgts.n.log 0 0.082741631
## .clusterid 1 0.080966421
## .clusterid.fctr 0 0.080966421
## D.chrs.pnct14.n.log 0 0.075911474
## D.terms.post.stop.n 1 0.075731613
## D.terms.post.stem.n 1 0.074990216
## cellular.fctr 0 0.073025939
## D.T.in. 0 0.071653919
## D.T.with 0 0.066087601
## D.terms.post.stop.n.log 0 0.062526435
## D.terms.post.stem.n.log 0 0.062220631
## D.wrds.unq.n.log 0 0.062220631
## D.chrs.pnct09.n.log 0 0.061919863
## D.ratio.wrds.stop.n.wrds.n 0 0.059547328
## carrier.fctr 0 0.058701911
## D.T.mint 0 0.055691207
## D.wrds.n.log 0 0.055678363
## D.chrs.n.log 0 0.055576665
## D.chrs.uppr.n.log 0 0.054491610
## prdl.my.fctr 0 0.054151937
## D.chrs.pnct28.n.log 0 0.052538378
## D.T.of 0 0.052259530
## D.chrs.pnct12.n.log 0 0.049684867
## D.chrs.pnct15.n.log 0 0.048611857
## startprice.dcm2.is9 0 0.047716416
## D.weight.post.stem.sum 0 0.047105442
## D.weight.sum 0 0.047105442
## D.weight.post.stop.sum 0 0.045009208
## color.fctr 0 0.044231522
## D.terms.n.stem.stop.Ratio 0 0.043562113
## startprice.dgt3.is9 0 0.043150483
## D.T.minor 0 0.042692360
## D.T.screen 0 0.040605274
## startprice.dgt2.is9 0 0.039616843
## D.chrs.pnct08.n.log 0 0.039606364
## D.T.condit 0 0.033248924
## D.T.light 0 0.032814746
## D.chrs.pnct13.n.log 0 0.032095777
## D.T.from 0 0.030653359
## D.chrs.pnct17.n.log 0 0.025087653
## D.chrs.pnct07.n.log 0 0.024330390
## D.chrs.pnct10.n.log 0 0.024107007
## D.T.or 0 0.022614420
## D.chrs.pnct03.n.log 0 0.021557732
## D.chrs.pnct18.n.log 0 0.021557732
## D.P.gold 0 0.021557732
## D.chrs.pnct11.n.log 0 0.020186238
## D.P.white 0 0.018597281
## D.T.good 0 0.016394675
## D.chrs.pnct16.n.log 0 0.015642041
## storage.fctr 0 0.014549906
## D.P.air 0 0.009203749
## startprice.dcm1.is9 0 0.008364097
## D.P.mini 0 0.006894908
## D.chrs.pnct01.n.log 0 0.004291712
## D.wrds.stop.n.log 0 0.003851183
## D.P.spacegray 0 0.003532788
## D.T.use 0 0.003025811
## D.weight.sum.stem.stop.Ratio 0 0.002381527
## .rnorm 0 0.001273806
## D.P.black 0 0.001181544
## D.ratio.weight.sum.wrds.n 0 0.001080251
## D.chrs.pnct02.n.log 0 NA
## D.chrs.pnct04.n.log 0 NA
## D.chrs.pnct19.n.log 0 NA
## D.chrs.pnct20.n.log 0 NA
## D.chrs.pnct21.n.log 0 NA
## D.chrs.pnct22.n.log 0 NA
## D.chrs.pnct23.n.log 0 NA
## D.chrs.pnct24.n.log 0 NA
## D.chrs.pnct25.n.log 0 NA
## D.chrs.pnct26.n.log 0 NA
## D.chrs.pnct27.n.log 0 NA
## D.chrs.pnct29.n.log 0 NA
## D.chrs.pnct30.n.log 0 NA
## D.P.http 0 NAprint(glb_feats_df <- orderBy(~-cor.y,
myfind_cor_features(feats_df=glb_feats_df, obs_df=glb_trnobs_df, rsp_var=glb_rsp_var,
nzv.freqCut=glb_nzv_freqCut, nzv.uniqueCut=glb_nzv_uniqueCut)))## [1] "cor(D.terms.post.stem.n.log, D.wrds.unq.n.log)=1.0000"
## [1] "cor(sold.fctr, D.terms.post.stem.n.log)=-0.0622"
## [1] "cor(sold.fctr, D.wrds.unq.n.log)=-0.0622"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified D.wrds.unq.n.log as highly correlated with
## D.terms.post.stem.n.log## [1] "cor(D.weight.post.stem.sum, D.weight.sum)=1.0000"
## [1] "cor(sold.fctr, D.weight.post.stem.sum)=-0.0471"
## [1] "cor(sold.fctr, D.weight.sum)=-0.0471"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified D.weight.sum as highly correlated with
## D.weight.post.stem.sum## [1] "cor(D.terms.post.stem.n.log, D.terms.post.stop.n.log)=0.9999"
## [1] "cor(sold.fctr, D.terms.post.stem.n.log)=-0.0622"
## [1] "cor(sold.fctr, D.terms.post.stop.n.log)=-0.0625"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified D.terms.post.stem.n.log as highly correlated
## with D.terms.post.stop.n.log## [1] "cor(D.chrs.n.log, D.chrs.uppr.n.log)=0.9996"
## [1] "cor(sold.fctr, D.chrs.n.log)=-0.0556"
## [1] "cor(sold.fctr, D.chrs.uppr.n.log)=-0.0545"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified D.chrs.uppr.n.log as highly correlated with
## D.chrs.n.log## [1] "cor(D.weight.post.stem.sum, D.weight.post.stop.sum)=0.9981"
## [1] "cor(sold.fctr, D.weight.post.stem.sum)=-0.0471"
## [1] "cor(sold.fctr, D.weight.post.stop.sum)=-0.0450"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified D.weight.post.stop.sum as highly correlated
## with D.weight.post.stem.sum## [1] "cor(D.terms.post.stop.n.log, D.wrds.n.log)=0.9956"
## [1] "cor(sold.fctr, D.terms.post.stop.n.log)=-0.0625"
## [1] "cor(sold.fctr, D.wrds.n.log)=-0.0557"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified D.wrds.n.log as highly correlated with
## D.terms.post.stop.n.log## [1] "cor(D.chrs.n.log, D.terms.post.stop.n.log)=0.9889"
## [1] "cor(sold.fctr, D.chrs.n.log)=-0.0556"
## [1] "cor(sold.fctr, D.terms.post.stop.n.log)=-0.0625"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified D.chrs.n.log as highly correlated with
## D.terms.post.stop.n.log## [1] "cor(D.ratio.wrds.stop.n.wrds.n, D.terms.post.stop.n.log)=-0.9745"
## [1] "cor(sold.fctr, D.ratio.wrds.stop.n.wrds.n)=0.0595"
## [1] "cor(sold.fctr, D.terms.post.stop.n.log)=-0.0625"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified D.ratio.wrds.stop.n.wrds.n as highly correlated
## with D.terms.post.stop.n.log## [1] "cor(sprice.d20nexp, sprice.log10)=-0.9316"
## [1] "cor(sold.fctr, sprice.d20nexp)=0.3980"
## [1] "cor(sold.fctr, sprice.log10)=-0.4694"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified sprice.d20nexp as highly correlated with
## sprice.log10## [1] "cor(D.terms.post.stop.n.log, D.weight.post.stem.sum)=0.9316"
## [1] "cor(sold.fctr, D.terms.post.stop.n.log)=-0.0625"
## [1] "cor(sold.fctr, D.weight.post.stem.sum)=-0.0471"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified D.weight.post.stem.sum as highly correlated
## with D.terms.post.stop.n.log## [1] "cor(D.T.X100, D.chrs.pnct05.n.log)=0.8813"
## [1] "cor(sold.fctr, D.T.X100)=-0.1114"
## [1] "cor(sold.fctr, D.chrs.pnct05.n.log)=-0.1202"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df
## = glb_trnobs_df, : Identified D.T.X100 as highly correlated with
## D.chrs.pnct05.n.log## [1] "cor(sprice.log10, sprice.root2)=0.8701"
## [1] "cor(sold.fctr, sprice.log10)=-0.4694"
## [1] "cor(sold.fctr, sprice.root2)=-0.5113"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified sprice.log10 as highly correlated with
## sprice.root2## [1] "cor(startprice.dcm1.is9, startprice.dcm2.is9)=0.8054"
## [1] "cor(sold.fctr, startprice.dcm1.is9)=-0.0084"
## [1] "cor(sold.fctr, startprice.dcm2.is9)=0.0477"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified startprice.dcm1.is9 as highly correlated with
## startprice.dcm2.is9## [1] "cor(D.chrs.pnct13.n.log, D.terms.post.stop.n.log)=0.7218"
## [1] "cor(sold.fctr, D.chrs.pnct13.n.log)=-0.0321"
## [1] "cor(sold.fctr, D.terms.post.stop.n.log)=-0.0625"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified D.chrs.pnct13.n.log as highly correlated with
## D.terms.post.stop.n.log## [1] "cor(carrier.fctr, cellular.fctr)=0.7131"
## [1] "cor(sold.fctr, carrier.fctr)=-0.0587"
## [1] "cor(sold.fctr, cellular.fctr)=-0.0730"## Warning in myfind_cor_features(feats_df = glb_feats_df, obs_df =
## glb_trnobs_df, : Identified carrier.fctr as highly correlated with
## cellular.fctr## id cor.y
## sold sold 1.000000000
## biddable biddable 0.549242842
## sprice.d20nexp sprice.d20nexp 0.397995133
## spdiff.cut.fctr spdiff.cut.fctr 0.290604966
## sprice.predict.diff sprice.predict.diff 0.271458109
## D.ratio.wrds.stop.n.wrds.n D.ratio.wrds.stop.n.wrds.n 0.059547328
## D.T.of D.T.of 0.052259530
## D.chrs.pnct15.n.log D.chrs.pnct15.n.log 0.048611857
## startprice.dcm2.is9 startprice.dcm2.is9 0.047716416
## D.terms.n.stem.stop.Ratio D.terms.n.stem.stop.Ratio 0.043562113
## D.T.screen D.T.screen 0.040605274
## D.chrs.pnct17.n.log D.chrs.pnct17.n.log 0.025087653
## D.chrs.pnct07.n.log D.chrs.pnct07.n.log 0.024330390
## D.P.white D.P.white 0.018597281
## D.T.good D.T.good 0.016394675
## D.chrs.pnct16.n.log D.chrs.pnct16.n.log 0.015642041
## D.chrs.pnct01.n.log D.chrs.pnct01.n.log 0.004291712
## D.wrds.stop.n.log D.wrds.stop.n.log 0.003851183
## D.P.spacegray D.P.spacegray 0.003532788
## D.T.use D.T.use 0.003025811
## D.weight.sum.stem.stop.Ratio D.weight.sum.stem.stop.Ratio 0.002381527
## D.ratio.weight.sum.wrds.n D.ratio.weight.sum.wrds.n -0.001080251
## D.P.black D.P.black -0.001181544
## .rnorm .rnorm -0.001273806
## D.P.mini D.P.mini -0.006894908
## startprice.dcm1.is9 startprice.dcm1.is9 -0.008364097
## D.P.air D.P.air -0.009203749
## storage.fctr storage.fctr -0.014549906
## D.chrs.pnct11.n.log D.chrs.pnct11.n.log -0.020186238
## D.P.gold D.P.gold -0.021557732
## D.chrs.pnct03.n.log D.chrs.pnct03.n.log -0.021557732
## D.chrs.pnct18.n.log D.chrs.pnct18.n.log -0.021557732
## D.T.or D.T.or -0.022614420
## D.chrs.pnct10.n.log D.chrs.pnct10.n.log -0.024107007
## D.T.from D.T.from -0.030653359
## D.chrs.pnct13.n.log D.chrs.pnct13.n.log -0.032095777
## D.T.light D.T.light -0.032814746
## D.T.condit D.T.condit -0.033248924
## D.chrs.pnct08.n.log D.chrs.pnct08.n.log -0.039606364
## startprice.dgt2.is9 startprice.dgt2.is9 -0.039616843
## D.T.minor D.T.minor -0.042692360
## startprice.dgt3.is9 startprice.dgt3.is9 -0.043150483
## color.fctr color.fctr -0.044231522
## D.weight.post.stop.sum D.weight.post.stop.sum -0.045009208
## D.weight.post.stem.sum D.weight.post.stem.sum -0.047105442
## D.weight.sum D.weight.sum -0.047105442
## D.chrs.pnct12.n.log D.chrs.pnct12.n.log -0.049684867
## D.chrs.pnct28.n.log D.chrs.pnct28.n.log -0.052538378
## prdl.my.fctr prdl.my.fctr -0.054151937
## D.chrs.uppr.n.log D.chrs.uppr.n.log -0.054491610
## D.chrs.n.log D.chrs.n.log -0.055576665
## D.wrds.n.log D.wrds.n.log -0.055678363
## D.T.mint D.T.mint -0.055691207
## carrier.fctr carrier.fctr -0.058701911
## D.chrs.pnct09.n.log D.chrs.pnct09.n.log -0.061919863
## D.terms.post.stem.n.log D.terms.post.stem.n.log -0.062220631
## D.wrds.unq.n.log D.wrds.unq.n.log -0.062220631
## D.terms.post.stop.n.log D.terms.post.stop.n.log -0.062526435
## D.T.with D.T.with -0.066087601
## D.T.in. D.T.in. -0.071653919
## cellular.fctr cellular.fctr -0.073025939
## D.terms.post.stem.n D.terms.post.stem.n -0.074990216
## D.terms.post.stop.n D.terms.post.stop.n -0.075731613
## D.chrs.pnct14.n.log D.chrs.pnct14.n.log -0.075911474
## .clusterid .clusterid -0.080966421
## .clusterid.fctr .clusterid.fctr -0.080966421
## D.dgts.n.log D.dgts.n.log -0.082741631
## D.T.cosmet D.T.cosmet -0.088480863
## D.chrs.pnct06.n.log D.chrs.pnct06.n.log -0.095888068
## D.T.X100 D.T.X100 -0.111398502
## D.chrs.pnct05.n.log D.chrs.pnct05.n.log -0.120173065
## D.T.hous D.T.hous -0.129274878
## startprice.dgt1.is9 startprice.dgt1.is9 -0.138476687
## condition.fctr condition.fctr -0.155283820
## UniqueID UniqueID -0.190626150
## startprice.log10.predict startprice.log10.predict -0.408722051
## startprice startprice -0.458981936
## sprice.log10 sprice.log10 -0.469398937
## sprice.root2 sprice.root2 -0.511275385
## D.P.http D.P.http NA
## D.chrs.pnct02.n.log D.chrs.pnct02.n.log NA
## D.chrs.pnct04.n.log D.chrs.pnct04.n.log NA
## D.chrs.pnct19.n.log D.chrs.pnct19.n.log NA
## D.chrs.pnct20.n.log D.chrs.pnct20.n.log NA
## D.chrs.pnct21.n.log D.chrs.pnct21.n.log NA
## D.chrs.pnct22.n.log D.chrs.pnct22.n.log NA
## D.chrs.pnct23.n.log D.chrs.pnct23.n.log NA
## D.chrs.pnct24.n.log D.chrs.pnct24.n.log NA
## D.chrs.pnct25.n.log D.chrs.pnct25.n.log NA
## D.chrs.pnct26.n.log D.chrs.pnct26.n.log NA
## D.chrs.pnct27.n.log D.chrs.pnct27.n.log NA
## D.chrs.pnct29.n.log D.chrs.pnct29.n.log NA
## D.chrs.pnct30.n.log D.chrs.pnct30.n.log NA
## exclude.as.feat cor.y.abs
## sold 1 1.000000000
## biddable 0 0.549242842
## sprice.d20nexp 0 0.397995133
## spdiff.cut.fctr 0 0.290604966
## sprice.predict.diff 1 0.271458109
## D.ratio.wrds.stop.n.wrds.n 0 0.059547328
## D.T.of 0 0.052259530
## D.chrs.pnct15.n.log 0 0.048611857
## startprice.dcm2.is9 0 0.047716416
## D.terms.n.stem.stop.Ratio 0 0.043562113
## D.T.screen 0 0.040605274
## D.chrs.pnct17.n.log 0 0.025087653
## D.chrs.pnct07.n.log 0 0.024330390
## D.P.white 0 0.018597281
## D.T.good 0 0.016394675
## D.chrs.pnct16.n.log 0 0.015642041
## D.chrs.pnct01.n.log 0 0.004291712
## D.wrds.stop.n.log 0 0.003851183
## D.P.spacegray 0 0.003532788
## D.T.use 0 0.003025811
## D.weight.sum.stem.stop.Ratio 0 0.002381527
## D.ratio.weight.sum.wrds.n 0 0.001080251
## D.P.black 0 0.001181544
## .rnorm 0 0.001273806
## D.P.mini 0 0.006894908
## startprice.dcm1.is9 0 0.008364097
## D.P.air 0 0.009203749
## storage.fctr 0 0.014549906
## D.chrs.pnct11.n.log 0 0.020186238
## D.P.gold 0 0.021557732
## D.chrs.pnct03.n.log 0 0.021557732
## D.chrs.pnct18.n.log 0 0.021557732
## D.T.or 0 0.022614420
## D.chrs.pnct10.n.log 0 0.024107007
## D.T.from 0 0.030653359
## D.chrs.pnct13.n.log 0 0.032095777
## D.T.light 0 0.032814746
## D.T.condit 0 0.033248924
## D.chrs.pnct08.n.log 0 0.039606364
## startprice.dgt2.is9 0 0.039616843
## D.T.minor 0 0.042692360
## startprice.dgt3.is9 0 0.043150483
## color.fctr 0 0.044231522
## D.weight.post.stop.sum 0 0.045009208
## D.weight.post.stem.sum 0 0.047105442
## D.weight.sum 0 0.047105442
## D.chrs.pnct12.n.log 0 0.049684867
## D.chrs.pnct28.n.log 0 0.052538378
## prdl.my.fctr 0 0.054151937
## D.chrs.uppr.n.log 0 0.054491610
## D.chrs.n.log 0 0.055576665
## D.wrds.n.log 0 0.055678363
## D.T.mint 0 0.055691207
## carrier.fctr 0 0.058701911
## D.chrs.pnct09.n.log 0 0.061919863
## D.terms.post.stem.n.log 0 0.062220631
## D.wrds.unq.n.log 0 0.062220631
## D.terms.post.stop.n.log 0 0.062526435
## D.T.with 0 0.066087601
## D.T.in. 0 0.071653919
## cellular.fctr 0 0.073025939
## D.terms.post.stem.n 1 0.074990216
## D.terms.post.stop.n 1 0.075731613
## D.chrs.pnct14.n.log 0 0.075911474
## .clusterid 1 0.080966421
## .clusterid.fctr 0 0.080966421
## D.dgts.n.log 0 0.082741631
## D.T.cosmet 0 0.088480863
## D.chrs.pnct06.n.log 0 0.095888068
## D.T.X100 0 0.111398502
## D.chrs.pnct05.n.log 0 0.120173065
## D.T.hous 0 0.129274878
## startprice.dgt1.is9 0 0.138476687
## condition.fctr 0 0.155283820
## UniqueID 1 0.190626150
## startprice.log10.predict 1 0.408722051
## startprice 1 0.458981936
## sprice.log10 0 0.469398937
## sprice.root2 0 0.511275385
## D.P.http 0 NA
## D.chrs.pnct02.n.log 0 NA
## D.chrs.pnct04.n.log 0 NA
## D.chrs.pnct19.n.log 0 NA
## D.chrs.pnct20.n.log 0 NA
## D.chrs.pnct21.n.log 0 NA
## D.chrs.pnct22.n.log 0 NA
## D.chrs.pnct23.n.log 0 NA
## D.chrs.pnct24.n.log 0 NA
## D.chrs.pnct25.n.log 0 NA
## D.chrs.pnct26.n.log 0 NA
## D.chrs.pnct27.n.log 0 NA
## D.chrs.pnct29.n.log 0 NA
## D.chrs.pnct30.n.log 0 NA
## cor.high.X freqRatio
## sold <NA> 1.163743
## biddable <NA> 1.215569
## sprice.d20nexp sprice.log10 2.807692
## spdiff.cut.fctr <NA> 1.668605
## sprice.predict.diff <NA> 1.000000
## D.ratio.wrds.stop.n.wrds.n D.terms.post.stop.n.log 18.067797
## D.T.of <NA> 145.000000
## D.chrs.pnct15.n.log <NA> 152.666667
## startprice.dcm2.is9 <NA> 2.195164
## D.terms.n.stem.stop.Ratio <NA> 85.142857
## D.T.screen <NA> 70.916667
## D.chrs.pnct17.n.log <NA> 1849.000000
## D.chrs.pnct07.n.log <NA> 107.529412
## D.P.white <NA> 230.125000
## D.T.good <NA> 95.166667
## D.chrs.pnct16.n.log <NA> 263.142857
## D.chrs.pnct01.n.log <NA> 52.705882
## D.wrds.stop.n.log <NA> 8.732484
## D.P.spacegray <NA> 461.500000
## D.T.use <NA> 86.631579
## D.weight.sum.stem.stop.Ratio <NA> 64.470588
## D.ratio.weight.sum.wrds.n <NA> 62.705882
## D.P.black <NA> 167.181818
## .rnorm <NA> 1.000000
## D.P.mini <NA> 96.315789
## startprice.dcm1.is9 startprice.dcm2.is9 1.616690
## D.P.air <NA> 122.266667
## storage.fctr <NA> 2.750742
## D.chrs.pnct11.n.log <NA> 9.382353
## D.P.gold <NA> 1849.000000
## D.chrs.pnct03.n.log <NA> 1849.000000
## D.chrs.pnct18.n.log <NA> 1849.000000
## D.T.or <NA> 87.950000
## D.chrs.pnct10.n.log <NA> 307.166667
## D.T.from <NA> 95.315789
## D.chrs.pnct13.n.log D.terms.post.stop.n.log 5.272374
## D.T.light <NA> 99.277778
## D.T.condit <NA> 36.829268
## D.chrs.pnct08.n.log <NA> 69.230769
## startprice.dgt2.is9 <NA> 6.905983
## D.T.minor <NA> 92.526316
## startprice.dgt3.is9 <NA> 461.500000
## color.fctr <NA> 1.592342
## D.weight.post.stop.sum D.weight.post.stem.sum 62.705882
## D.weight.post.stem.sum D.terms.post.stop.n.log 62.705882
## D.weight.sum D.weight.post.stem.sum 62.705882
## D.chrs.pnct12.n.log <NA> 28.467742
## D.chrs.pnct28.n.log <NA> 461.000000
## prdl.my.fctr <NA> 1.032491
## D.chrs.uppr.n.log D.chrs.n.log 15.661765
## D.chrs.n.log D.terms.post.stop.n.log 13.455696
## D.wrds.n.log D.terms.post.stop.n.log 11.430108
## D.T.mint <NA> 94.894737
## carrier.fctr cellular.fctr 3.184971
## D.chrs.pnct09.n.log <NA> 306.833333
## D.terms.post.stem.n.log D.terms.post.stop.n.log 11.714286
## D.wrds.unq.n.log D.terms.post.stem.n.log 11.714286
## D.terms.post.stop.n.log <NA> 13.325000
## D.T.with <NA> 46.888889
## D.T.in. <NA> 33.521739
## cellular.fctr <NA> 2.103053
## D.terms.post.stem.n <NA> 11.714286
## D.terms.post.stop.n <NA> 13.325000
## D.chrs.pnct14.n.log <NA> 35.176471
## .clusterid <NA> 10.959732
## .clusterid.fctr <NA> 10.959732
## D.dgts.n.log <NA> 34.340000
## D.T.cosmet <NA> 88.400000
## D.chrs.pnct06.n.log <NA> 60.466667
## D.T.X100 D.chrs.pnct05.n.log 180.500000
## D.chrs.pnct05.n.log <NA> 39.217391
## D.T.hous <NA> 81.681818
## startprice.dgt1.is9 <NA> 2.042763
## condition.fctr <NA> 4.010453
## UniqueID <NA> 1.000000
## startprice.log10.predict <NA> 1.230769
## startprice <NA> 2.807692
## sprice.log10 sprice.root2 2.807692
## sprice.root2 <NA> 2.807692
## D.P.http <NA> 0.000000
## D.chrs.pnct02.n.log <NA> 0.000000
## D.chrs.pnct04.n.log <NA> 0.000000
## D.chrs.pnct19.n.log <NA> 0.000000
## D.chrs.pnct20.n.log <NA> 0.000000
## D.chrs.pnct21.n.log <NA> 0.000000
## D.chrs.pnct22.n.log <NA> 0.000000
## D.chrs.pnct23.n.log <NA> 0.000000
## D.chrs.pnct24.n.log <NA> 0.000000
## D.chrs.pnct25.n.log <NA> 0.000000
## D.chrs.pnct26.n.log <NA> 0.000000
## D.chrs.pnct27.n.log <NA> 0.000000
## D.chrs.pnct29.n.log <NA> 0.000000
## D.chrs.pnct30.n.log <NA> 0.000000
## percentUnique zeroVar nzv is.cor.y.abs.low
## sold 0.10810811 FALSE FALSE FALSE
## biddable 0.10810811 FALSE FALSE FALSE
## sprice.d20nexp 30.21621622 FALSE FALSE FALSE
## spdiff.cut.fctr 0.43243243 FALSE FALSE FALSE
## sprice.predict.diff 96.10810811 FALSE FALSE FALSE
## D.ratio.wrds.stop.n.wrds.n 4.27027027 FALSE FALSE FALSE
## D.T.of 1.13513514 FALSE TRUE FALSE
## D.chrs.pnct15.n.log 0.16216216 FALSE TRUE FALSE
## startprice.dcm2.is9 0.10810811 FALSE FALSE FALSE
## D.terms.n.stem.stop.Ratio 0.64864865 FALSE TRUE FALSE
## D.T.screen 1.13513514 FALSE TRUE FALSE
## D.chrs.pnct17.n.log 0.10810811 FALSE TRUE FALSE
## D.chrs.pnct07.n.log 0.16216216 FALSE TRUE FALSE
## D.P.white 0.16216216 FALSE TRUE FALSE
## D.T.good 1.18918919 FALSE TRUE FALSE
## D.chrs.pnct16.n.log 0.16216216 FALSE TRUE FALSE
## D.chrs.pnct01.n.log 0.32432432 FALSE TRUE FALSE
## D.wrds.stop.n.log 0.54054054 FALSE FALSE FALSE
## D.P.spacegray 0.10810811 FALSE TRUE FALSE
## D.T.use 1.51351351 FALSE TRUE FALSE
## D.weight.sum.stem.stop.Ratio 33.56756757 FALSE FALSE FALSE
## D.ratio.weight.sum.wrds.n 34.81081081 FALSE FALSE TRUE
## D.P.black 0.10810811 FALSE TRUE TRUE
## .rnorm 100.00000000 FALSE FALSE FALSE
## D.P.mini 0.16216216 FALSE TRUE FALSE
## startprice.dcm1.is9 0.10810811 FALSE FALSE FALSE
## D.P.air 0.16216216 FALSE TRUE FALSE
## storage.fctr 0.27027027 FALSE FALSE FALSE
## D.chrs.pnct11.n.log 0.37837838 FALSE FALSE FALSE
## D.P.gold 0.10810811 FALSE TRUE FALSE
## D.chrs.pnct03.n.log 0.10810811 FALSE TRUE FALSE
## D.chrs.pnct18.n.log 0.10810811 FALSE TRUE FALSE
## D.T.or 1.08108108 FALSE TRUE FALSE
## D.chrs.pnct10.n.log 0.16216216 FALSE TRUE FALSE
## D.T.from 0.70270270 FALSE TRUE FALSE
## D.chrs.pnct13.n.log 0.48648649 FALSE FALSE FALSE
## D.T.light 0.91891892 FALSE TRUE FALSE
## D.T.condit 1.35135135 FALSE TRUE FALSE
## D.chrs.pnct08.n.log 0.21621622 FALSE TRUE FALSE
## startprice.dgt2.is9 0.10810811 FALSE FALSE FALSE
## D.T.minor 0.97297297 FALSE TRUE FALSE
## startprice.dgt3.is9 0.10810811 FALSE TRUE FALSE
## color.fctr 0.27027027 FALSE FALSE FALSE
## D.weight.post.stop.sum 34.75675676 FALSE FALSE FALSE
## D.weight.post.stem.sum 34.64864865 FALSE FALSE FALSE
## D.weight.sum 34.64864865 FALSE FALSE FALSE
## D.chrs.pnct12.n.log 0.21621622 FALSE TRUE FALSE
## D.chrs.pnct28.n.log 0.16216216 FALSE TRUE FALSE
## prdl.my.fctr 0.54054054 FALSE FALSE FALSE
## D.chrs.uppr.n.log 4.48648649 FALSE FALSE FALSE
## D.chrs.n.log 5.35135135 FALSE FALSE FALSE
## D.wrds.n.log 1.29729730 FALSE FALSE FALSE
## D.T.mint 1.02702703 FALSE TRUE FALSE
## carrier.fctr 0.37837838 FALSE FALSE FALSE
## D.chrs.pnct09.n.log 0.21621622 FALSE TRUE FALSE
## D.terms.post.stem.n.log 1.13513514 FALSE FALSE FALSE
## D.wrds.unq.n.log 1.13513514 FALSE FALSE FALSE
## D.terms.post.stop.n.log 1.13513514 FALSE FALSE FALSE
## D.T.with 1.02702703 FALSE TRUE FALSE
## D.T.in. 1.35135135 FALSE TRUE FALSE
## cellular.fctr 0.16216216 FALSE FALSE FALSE
## D.terms.post.stem.n 1.13513514 FALSE FALSE FALSE
## D.terms.post.stop.n 1.13513514 FALSE FALSE FALSE
## D.chrs.pnct14.n.log 0.27027027 FALSE TRUE FALSE
## .clusterid 0.16216216 FALSE FALSE FALSE
## .clusterid.fctr 0.16216216 FALSE FALSE FALSE
## D.dgts.n.log 0.59459459 FALSE TRUE FALSE
## D.T.cosmet 0.81081081 FALSE TRUE FALSE
## D.chrs.pnct06.n.log 0.16216216 FALSE TRUE FALSE
## D.T.X100 0.64864865 FALSE TRUE FALSE
## D.chrs.pnct05.n.log 0.10810811 FALSE TRUE FALSE
## D.T.hous 0.54054054 FALSE TRUE FALSE
## startprice.dgt1.is9 0.10810811 FALSE FALSE FALSE
## condition.fctr 0.32432432 FALSE FALSE FALSE
## UniqueID 100.00000000 FALSE FALSE FALSE
## startprice.log10.predict 87.94594595 FALSE FALSE FALSE
## startprice 30.21621622 FALSE FALSE FALSE
## sprice.log10 30.21621622 FALSE FALSE FALSE
## sprice.root2 30.21621622 FALSE FALSE FALSE
## D.P.http 0.05405405 TRUE TRUE NA
## D.chrs.pnct02.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct04.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct19.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct20.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct21.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct22.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct23.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct24.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct25.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct26.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct27.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct29.n.log 0.05405405 TRUE TRUE NA
## D.chrs.pnct30.n.log 0.05405405 TRUE TRUE NAplt_feats_df <- glb_feats_df
print(myplot_scatter(plt_feats_df, "percentUnique", "freqRatio",
colorcol_name="nzv", jitter=TRUE) +
#geom_point(aes(shape=nzv)) +
geom_point() +
xlim(-5, 25) +
geom_hline(yintercept=glb_nzv_freqCut) +
geom_vline(xintercept=glb_nzv_uniqueCut))## Warning in myplot_scatter(plt_feats_df, "percentUnique", "freqRatio",
## colorcol_name = "nzv", : converting nzv to class:factor## Warning: Removed 13 rows containing missing values (geom_point).## Warning: Removed 13 rows containing missing values (geom_point).## Warning: Removed 13 rows containing missing values (geom_point).
print(subset(glb_feats_df, nzv))## id cor.y
## D.T.of D.T.of 0.052259530
## D.chrs.pnct15.n.log D.chrs.pnct15.n.log 0.048611857
## D.terms.n.stem.stop.Ratio D.terms.n.stem.stop.Ratio 0.043562113
## D.T.screen D.T.screen 0.040605274
## D.chrs.pnct17.n.log D.chrs.pnct17.n.log 0.025087653
## D.chrs.pnct07.n.log D.chrs.pnct07.n.log 0.024330390
## D.P.white D.P.white 0.018597281
## D.T.good D.T.good 0.016394675
## D.chrs.pnct16.n.log D.chrs.pnct16.n.log 0.015642041
## D.chrs.pnct01.n.log D.chrs.pnct01.n.log 0.004291712
## D.P.spacegray D.P.spacegray 0.003532788
## D.T.use D.T.use 0.003025811
## D.P.black D.P.black -0.001181544
## D.P.mini D.P.mini -0.006894908
## D.P.air D.P.air -0.009203749
## D.P.gold D.P.gold -0.021557732
## D.chrs.pnct03.n.log D.chrs.pnct03.n.log -0.021557732
## D.chrs.pnct18.n.log D.chrs.pnct18.n.log -0.021557732
## D.T.or D.T.or -0.022614420
## D.chrs.pnct10.n.log D.chrs.pnct10.n.log -0.024107007
## D.T.from D.T.from -0.030653359
## D.T.light D.T.light -0.032814746
## D.T.condit D.T.condit -0.033248924
## D.chrs.pnct08.n.log D.chrs.pnct08.n.log -0.039606364
## D.T.minor D.T.minor -0.042692360
## startprice.dgt3.is9 startprice.dgt3.is9 -0.043150483
## D.chrs.pnct12.n.log D.chrs.pnct12.n.log -0.049684867
## D.chrs.pnct28.n.log D.chrs.pnct28.n.log -0.052538378
## D.T.mint D.T.mint -0.055691207
## D.chrs.pnct09.n.log D.chrs.pnct09.n.log -0.061919863
## D.T.with D.T.with -0.066087601
## D.T.in. D.T.in. -0.071653919
## D.chrs.pnct14.n.log D.chrs.pnct14.n.log -0.075911474
## D.dgts.n.log D.dgts.n.log -0.082741631
## D.T.cosmet D.T.cosmet -0.088480863
## D.chrs.pnct06.n.log D.chrs.pnct06.n.log -0.095888068
## D.T.X100 D.T.X100 -0.111398502
## D.chrs.pnct05.n.log D.chrs.pnct05.n.log -0.120173065
## D.T.hous D.T.hous -0.129274878
## D.P.http D.P.http NA
## D.chrs.pnct02.n.log D.chrs.pnct02.n.log NA
## D.chrs.pnct04.n.log D.chrs.pnct04.n.log NA
## D.chrs.pnct19.n.log D.chrs.pnct19.n.log NA
## D.chrs.pnct20.n.log D.chrs.pnct20.n.log NA
## D.chrs.pnct21.n.log D.chrs.pnct21.n.log NA
## D.chrs.pnct22.n.log D.chrs.pnct22.n.log NA
## D.chrs.pnct23.n.log D.chrs.pnct23.n.log NA
## D.chrs.pnct24.n.log D.chrs.pnct24.n.log NA
## D.chrs.pnct25.n.log D.chrs.pnct25.n.log NA
## D.chrs.pnct26.n.log D.chrs.pnct26.n.log NA
## D.chrs.pnct27.n.log D.chrs.pnct27.n.log NA
## D.chrs.pnct29.n.log D.chrs.pnct29.n.log NA
## D.chrs.pnct30.n.log D.chrs.pnct30.n.log NA
## exclude.as.feat cor.y.abs cor.high.X
## D.T.of 0 0.052259530 <NA>
## D.chrs.pnct15.n.log 0 0.048611857 <NA>
## D.terms.n.stem.stop.Ratio 0 0.043562113 <NA>
## D.T.screen 0 0.040605274 <NA>
## D.chrs.pnct17.n.log 0 0.025087653 <NA>
## D.chrs.pnct07.n.log 0 0.024330390 <NA>
## D.P.white 0 0.018597281 <NA>
## D.T.good 0 0.016394675 <NA>
## D.chrs.pnct16.n.log 0 0.015642041 <NA>
## D.chrs.pnct01.n.log 0 0.004291712 <NA>
## D.P.spacegray 0 0.003532788 <NA>
## D.T.use 0 0.003025811 <NA>
## D.P.black 0 0.001181544 <NA>
## D.P.mini 0 0.006894908 <NA>
## D.P.air 0 0.009203749 <NA>
## D.P.gold 0 0.021557732 <NA>
## D.chrs.pnct03.n.log 0 0.021557732 <NA>
## D.chrs.pnct18.n.log 0 0.021557732 <NA>
## D.T.or 0 0.022614420 <NA>
## D.chrs.pnct10.n.log 0 0.024107007 <NA>
## D.T.from 0 0.030653359 <NA>
## D.T.light 0 0.032814746 <NA>
## D.T.condit 0 0.033248924 <NA>
## D.chrs.pnct08.n.log 0 0.039606364 <NA>
## D.T.minor 0 0.042692360 <NA>
## startprice.dgt3.is9 0 0.043150483 <NA>
## D.chrs.pnct12.n.log 0 0.049684867 <NA>
## D.chrs.pnct28.n.log 0 0.052538378 <NA>
## D.T.mint 0 0.055691207 <NA>
## D.chrs.pnct09.n.log 0 0.061919863 <NA>
## D.T.with 0 0.066087601 <NA>
## D.T.in. 0 0.071653919 <NA>
## D.chrs.pnct14.n.log 0 0.075911474 <NA>
## D.dgts.n.log 0 0.082741631 <NA>
## D.T.cosmet 0 0.088480863 <NA>
## D.chrs.pnct06.n.log 0 0.095888068 <NA>
## D.T.X100 0 0.111398502 D.chrs.pnct05.n.log
## D.chrs.pnct05.n.log 0 0.120173065 <NA>
## D.T.hous 0 0.129274878 <NA>
## D.P.http 0 NA <NA>
## D.chrs.pnct02.n.log 0 NA <NA>
## D.chrs.pnct04.n.log 0 NA <NA>
## D.chrs.pnct19.n.log 0 NA <NA>
## D.chrs.pnct20.n.log 0 NA <NA>
## D.chrs.pnct21.n.log 0 NA <NA>
## D.chrs.pnct22.n.log 0 NA <NA>
## D.chrs.pnct23.n.log 0 NA <NA>
## D.chrs.pnct24.n.log 0 NA <NA>
## D.chrs.pnct25.n.log 0 NA <NA>
## D.chrs.pnct26.n.log 0 NA <NA>
## D.chrs.pnct27.n.log 0 NA <NA>
## D.chrs.pnct29.n.log 0 NA <NA>
## D.chrs.pnct30.n.log 0 NA <NA>
## freqRatio percentUnique zeroVar nzv
## D.T.of 145.00000 1.13513514 FALSE TRUE
## D.chrs.pnct15.n.log 152.66667 0.16216216 FALSE TRUE
## D.terms.n.stem.stop.Ratio 85.14286 0.64864865 FALSE TRUE
## D.T.screen 70.91667 1.13513514 FALSE TRUE
## D.chrs.pnct17.n.log 1849.00000 0.10810811 FALSE TRUE
## D.chrs.pnct07.n.log 107.52941 0.16216216 FALSE TRUE
## D.P.white 230.12500 0.16216216 FALSE TRUE
## D.T.good 95.16667 1.18918919 FALSE TRUE
## D.chrs.pnct16.n.log 263.14286 0.16216216 FALSE TRUE
## D.chrs.pnct01.n.log 52.70588 0.32432432 FALSE TRUE
## D.P.spacegray 461.50000 0.10810811 FALSE TRUE
## D.T.use 86.63158 1.51351351 FALSE TRUE
## D.P.black 167.18182 0.10810811 FALSE TRUE
## D.P.mini 96.31579 0.16216216 FALSE TRUE
## D.P.air 122.26667 0.16216216 FALSE TRUE
## D.P.gold 1849.00000 0.10810811 FALSE TRUE
## D.chrs.pnct03.n.log 1849.00000 0.10810811 FALSE TRUE
## D.chrs.pnct18.n.log 1849.00000 0.10810811 FALSE TRUE
## D.T.or 87.95000 1.08108108 FALSE TRUE
## D.chrs.pnct10.n.log 307.16667 0.16216216 FALSE TRUE
## D.T.from 95.31579 0.70270270 FALSE TRUE
## D.T.light 99.27778 0.91891892 FALSE TRUE
## D.T.condit 36.82927 1.35135135 FALSE TRUE
## D.chrs.pnct08.n.log 69.23077 0.21621622 FALSE TRUE
## D.T.minor 92.52632 0.97297297 FALSE TRUE
## startprice.dgt3.is9 461.50000 0.10810811 FALSE TRUE
## D.chrs.pnct12.n.log 28.46774 0.21621622 FALSE TRUE
## D.chrs.pnct28.n.log 461.00000 0.16216216 FALSE TRUE
## D.T.mint 94.89474 1.02702703 FALSE TRUE
## D.chrs.pnct09.n.log 306.83333 0.21621622 FALSE TRUE
## D.T.with 46.88889 1.02702703 FALSE TRUE
## D.T.in. 33.52174 1.35135135 FALSE TRUE
## D.chrs.pnct14.n.log 35.17647 0.27027027 FALSE TRUE
## D.dgts.n.log 34.34000 0.59459459 FALSE TRUE
## D.T.cosmet 88.40000 0.81081081 FALSE TRUE
## D.chrs.pnct06.n.log 60.46667 0.16216216 FALSE TRUE
## D.T.X100 180.50000 0.64864865 FALSE TRUE
## D.chrs.pnct05.n.log 39.21739 0.10810811 FALSE TRUE
## D.T.hous 81.68182 0.54054054 FALSE TRUE
## D.P.http 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct02.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct04.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct19.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct20.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct21.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct22.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct23.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct24.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct25.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct26.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct27.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct29.n.log 0.00000 0.05405405 TRUE TRUE
## D.chrs.pnct30.n.log 0.00000 0.05405405 TRUE TRUE
## is.cor.y.abs.low
## D.T.of FALSE
## D.chrs.pnct15.n.log FALSE
## D.terms.n.stem.stop.Ratio FALSE
## D.T.screen FALSE
## D.chrs.pnct17.n.log FALSE
## D.chrs.pnct07.n.log FALSE
## D.P.white FALSE
## D.T.good FALSE
## D.chrs.pnct16.n.log FALSE
## D.chrs.pnct01.n.log FALSE
## D.P.spacegray FALSE
## D.T.use FALSE
## D.P.black TRUE
## D.P.mini FALSE
## D.P.air FALSE
## D.P.gold FALSE
## D.chrs.pnct03.n.log FALSE
## D.chrs.pnct18.n.log FALSE
## D.T.or FALSE
## D.chrs.pnct10.n.log FALSE
## D.T.from FALSE
## D.T.light FALSE
## D.T.condit FALSE
## D.chrs.pnct08.n.log FALSE
## D.T.minor FALSE
## startprice.dgt3.is9 FALSE
## D.chrs.pnct12.n.log FALSE
## D.chrs.pnct28.n.log FALSE
## D.T.mint FALSE
## D.chrs.pnct09.n.log FALSE
## D.T.with FALSE
## D.T.in. FALSE
## D.chrs.pnct14.n.log FALSE
## D.dgts.n.log FALSE
## D.T.cosmet FALSE
## D.chrs.pnct06.n.log FALSE
## D.T.X100 FALSE
## D.chrs.pnct05.n.log FALSE
## D.T.hous FALSE
## D.P.http NA
## D.chrs.pnct02.n.log NA
## D.chrs.pnct04.n.log NA
## D.chrs.pnct19.n.log NA
## D.chrs.pnct20.n.log NA
## D.chrs.pnct21.n.log NA
## D.chrs.pnct22.n.log NA
## D.chrs.pnct23.n.log NA
## D.chrs.pnct24.n.log NA
## D.chrs.pnct25.n.log NA
## D.chrs.pnct26.n.log NA
## D.chrs.pnct27.n.log NA
## D.chrs.pnct29.n.log NA
## D.chrs.pnct30.n.log NAtmp_allobs_df <-
glb_allobs_df[, union(setdiff(names(glb_allobs_df), subset(glb_feats_df, nzv)$id),
glb_cluster_entropy_var)]
glb_trnobs_df <- subset(tmp_allobs_df, .src == "Train")
glb_newobs_df <- subset(tmp_allobs_df, .src == "Test")
glb_feats_df$interaction.feat <- NA
for (feat in names(glb_interaction_only_feats_lst))
glb_feats_df[glb_feats_df$id %in% feat, "interaction.feat"] <-
glb_interaction_only_feats_lst[[feat]]
#stop(here"); glb_to_sav(); glb_allobs_df <- sav_allobs_df
indep_vars <- subset(glb_feats_df, !nzv & (exclude.as.feat != 1))[, "id"]
numeric_indep_vars <- indep_vars[!grepl(".fctr", indep_vars, fixed=TRUE)]
glb_feats_df$shapiro.test.p.value <- NA
glb_feats_df[glb_feats_df$id %in% numeric_indep_vars, "shapiro.test.p.value"] <-
sapply(numeric_indep_vars, function(var) shapiro.test(glb_trnobs_df[, var])$p.value)
not_nrml_feats_df <- glb_feats_df %>%
subset(!is.na(shapiro.test.p.value)) %>%
subset((shapiro.test.p.value < 0.05) || (id == ".rnorm")) %>%
arrange(shapiro.test.p.value)
row.names(not_nrml_feats_df) <- not_nrml_feats_df$id
#plt_trnobs_df <- glb_trnobs_df[, c("D.npnct05.log", ".rnorm")]
plt_trnobs_df <- glb_trnobs_df[, c(union(not_nrml_feats_df$id[1:min(5, nrow(not_nrml_feats_df))],
".rnorm"), glb_cluster_entropy_var)]
print(myplot_violin(plt_trnobs_df, setdiff(names(plt_trnobs_df), glb_cluster_entropy_var),
xcol_name = glb_cluster_entropy_var) +
facet_wrap(~variable, scales="free"))
#myplot_histogram(plt_trnobs_df, "D.npnct11.log", fill_col_name="sold", show_stats = TRUE)
myadjust_interaction_feats <- function(vars_vctr) {
for (feat in subset(glb_feats_df, !is.na(interaction.feat))$id)
if (feat %in% vars_vctr)
vars_vctr <- union(setdiff(vars_vctr, feat),
paste0(glb_feats_df[glb_feats_df$id == feat, "interaction.feat"], ":",
feat))
return(vars_vctr)
}
myrun_rfe <- function(obs_df, indep_vars, sizes = NULL) {
rfe_obs_df <- myget_vectorized_obs_df(obs_df, glb_rsp_var, indep_vars)
predictors_vctr <- setdiff(names(rfe_obs_df), glb_rsp_var)
if (glb_is_regression) rfeFuncs <- lmFuncs else {
rfeFuncs <- ldaFuncs
# Delete non-variant columns
predictors_unqLen <- sapply(predictors_vctr, function(col)
length(unique(rfe_obs_df[, col])))
predictors_vctr <- predictors_vctr[predictors_unqLen > 1]
# Delete freqRatio >= 291
# plagiarized from caret:::nzv
predictors_freqRatio <-
apply(rfe_obs_df[, predictors_vctr], 2, function(data) {
t <- table(data[!is.na(data)])
if (length(t) <= 1) {
return(0)
}
w <- which.max(t)
return(max(t, na.rm = TRUE)/max(t[-w], na.rm = TRUE))
})
predictors_vctr <- predictors_vctr[predictors_freqRatio < 172]
}
if (is.null(sizes))
sizes <- tail(2 ^ (1:as.integer(log2(length(predictors_vctr)))), 5)
rfe_control <- rfeControl(functions = rfeFuncs, method = "repeatedcv",
number = glb_rcv_n_folds, repeats = glb_rcv_n_repeats,
verbose = TRUE, returnResamp = "all",
seeds = mygen_seeds(seeds_lst_len =
(glb_rcv_n_folds * glb_rcv_n_repeats) + 1,
seeds_elmnt_lst_len = (length(sizes) + 1))
, allowParallel = FALSE
)
set.seed(113)
rfe_results <- rfe(rfe_obs_df[, predictors_vctr],
rfe_obs_df[, glb_rsp_var],
sizes = sizes,
# metric = unlist(strsplit(glbMdlMetricsEval, "[.]"))[2],
# maximize = ifelse(unlist(strsplit(glbMdlMetricsEval, "[.]"))[1] == "max",
# TRUE, FALSE),
rfeControl = rfe_control)
print(rfe_results)
print(predictors(rfe_results))
# print(plot(rfe_results, type=c("g", "o")))
# print(plot(rfe_results))
print(ggplot(rfe_results))
return(rfe_results)
}
#stop(here"); glb_to_sav()
# shd .clusterid.fctr be excluded from this ? or include encoding of glb_category_var:.clusterid.fctr ?
indep_vars <- myadjust_interaction_feats(subset(glb_feats_df,
!nzv & (exclude.as.feat != 1))$id)
rfe_fit_results <- myrun_rfe(obs_df = glb_fitobs_df, indep_vars = indep_vars,
sizes = glb_rfe_fit_sizes)## +(rfe) fit Fold1.Rep1 size: 70## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold1.Rep1 size: 70
## +(rfe) imp Fold1.Rep1
## -(rfe) imp Fold1.Rep1
## +(rfe) fit Fold1.Rep1 size: 64## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold1.Rep1 size: 64
## +(rfe) fit Fold1.Rep1 size: 32
## -(rfe) fit Fold1.Rep1 size: 32
## +(rfe) fit Fold1.Rep1 size: 16
## -(rfe) fit Fold1.Rep1 size: 16
## +(rfe) fit Fold1.Rep1 size: 8
## -(rfe) fit Fold1.Rep1 size: 8
## +(rfe) fit Fold1.Rep1 size: 4
## -(rfe) fit Fold1.Rep1 size: 4
## +(rfe) fit Fold2.Rep1 size: 70## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold2.Rep1 size: 70
## +(rfe) imp Fold2.Rep1
## -(rfe) imp Fold2.Rep1
## +(rfe) fit Fold2.Rep1 size: 64## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold2.Rep1 size: 64
## +(rfe) fit Fold2.Rep1 size: 32
## -(rfe) fit Fold2.Rep1 size: 32
## +(rfe) fit Fold2.Rep1 size: 16
## -(rfe) fit Fold2.Rep1 size: 16
## +(rfe) fit Fold2.Rep1 size: 8
## -(rfe) fit Fold2.Rep1 size: 8
## +(rfe) fit Fold2.Rep1 size: 4
## -(rfe) fit Fold2.Rep1 size: 4
## +(rfe) fit Fold3.Rep1 size: 70## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold3.Rep1 size: 70
## +(rfe) imp Fold3.Rep1
## -(rfe) imp Fold3.Rep1
## +(rfe) fit Fold3.Rep1 size: 64## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold3.Rep1 size: 64
## +(rfe) fit Fold3.Rep1 size: 32
## -(rfe) fit Fold3.Rep1 size: 32
## +(rfe) fit Fold3.Rep1 size: 16
## -(rfe) fit Fold3.Rep1 size: 16
## +(rfe) fit Fold3.Rep1 size: 8
## -(rfe) fit Fold3.Rep1 size: 8
## +(rfe) fit Fold3.Rep1 size: 4
## -(rfe) fit Fold3.Rep1 size: 4
## +(rfe) fit Fold1.Rep2 size: 70## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold1.Rep2 size: 70
## +(rfe) imp Fold1.Rep2
## -(rfe) imp Fold1.Rep2
## +(rfe) fit Fold1.Rep2 size: 64## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold1.Rep2 size: 64
## +(rfe) fit Fold1.Rep2 size: 32
## -(rfe) fit Fold1.Rep2 size: 32
## +(rfe) fit Fold1.Rep2 size: 16
## -(rfe) fit Fold1.Rep2 size: 16
## +(rfe) fit Fold1.Rep2 size: 8
## -(rfe) fit Fold1.Rep2 size: 8
## +(rfe) fit Fold1.Rep2 size: 4
## -(rfe) fit Fold1.Rep2 size: 4
## +(rfe) fit Fold2.Rep2 size: 70## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold2.Rep2 size: 70
## +(rfe) imp Fold2.Rep2
## -(rfe) imp Fold2.Rep2
## +(rfe) fit Fold2.Rep2 size: 64## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold2.Rep2 size: 64
## +(rfe) fit Fold2.Rep2 size: 32
## -(rfe) fit Fold2.Rep2 size: 32
## +(rfe) fit Fold2.Rep2 size: 16
## -(rfe) fit Fold2.Rep2 size: 16
## +(rfe) fit Fold2.Rep2 size: 8
## -(rfe) fit Fold2.Rep2 size: 8
## +(rfe) fit Fold2.Rep2 size: 4
## -(rfe) fit Fold2.Rep2 size: 4
## +(rfe) fit Fold3.Rep2 size: 70## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold3.Rep2 size: 70
## +(rfe) imp Fold3.Rep2
## -(rfe) imp Fold3.Rep2
## +(rfe) fit Fold3.Rep2 size: 64## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold3.Rep2 size: 64
## +(rfe) fit Fold3.Rep2 size: 32
## -(rfe) fit Fold3.Rep2 size: 32
## +(rfe) fit Fold3.Rep2 size: 16
## -(rfe) fit Fold3.Rep2 size: 16
## +(rfe) fit Fold3.Rep2 size: 8
## -(rfe) fit Fold3.Rep2 size: 8
## +(rfe) fit Fold3.Rep2 size: 4
## -(rfe) fit Fold3.Rep2 size: 4
## +(rfe) fit Fold1.Rep3 size: 70## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold1.Rep3 size: 70
## +(rfe) imp Fold1.Rep3
## -(rfe) imp Fold1.Rep3
## +(rfe) fit Fold1.Rep3 size: 64## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold1.Rep3 size: 64
## +(rfe) fit Fold1.Rep3 size: 32
## -(rfe) fit Fold1.Rep3 size: 32
## +(rfe) fit Fold1.Rep3 size: 16
## -(rfe) fit Fold1.Rep3 size: 16
## +(rfe) fit Fold1.Rep3 size: 8
## -(rfe) fit Fold1.Rep3 size: 8
## +(rfe) fit Fold1.Rep3 size: 4
## -(rfe) fit Fold1.Rep3 size: 4
## +(rfe) fit Fold2.Rep3 size: 70## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold2.Rep3 size: 70
## +(rfe) imp Fold2.Rep3
## -(rfe) imp Fold2.Rep3
## +(rfe) fit Fold2.Rep3 size: 64## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold2.Rep3 size: 64
## +(rfe) fit Fold2.Rep3 size: 32
## -(rfe) fit Fold2.Rep3 size: 32
## +(rfe) fit Fold2.Rep3 size: 16
## -(rfe) fit Fold2.Rep3 size: 16
## +(rfe) fit Fold2.Rep3 size: 8
## -(rfe) fit Fold2.Rep3 size: 8
## +(rfe) fit Fold2.Rep3 size: 4
## -(rfe) fit Fold2.Rep3 size: 4
## +(rfe) fit Fold3.Rep3 size: 70## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold3.Rep3 size: 70
## +(rfe) imp Fold3.Rep3
## -(rfe) imp Fold3.Rep3
## +(rfe) fit Fold3.Rep3 size: 64## Warning in lda.default(x, grouping, ...): variables are collinear## -(rfe) fit Fold3.Rep3 size: 64
## +(rfe) fit Fold3.Rep3 size: 32
## -(rfe) fit Fold3.Rep3 size: 32
## +(rfe) fit Fold3.Rep3 size: 16
## -(rfe) fit Fold3.Rep3 size: 16
## +(rfe) fit Fold3.Rep3 size: 8
## -(rfe) fit Fold3.Rep3 size: 8
## +(rfe) fit Fold3.Rep3 size: 4
## -(rfe) fit Fold3.Rep3 size: 4## Warning in lda.default(x, grouping, ...): variables are collinear##
## Recursive feature selection
##
## Outer resampling method: Cross-Validated (3 fold, repeated 3 times)
##
## Resampling performance over subset size:
##
## Variables Accuracy Kappa AccuracySD KappaSD Selected
## 4 0.7951 0.5866 0.01779 0.03553
## 8 0.7954 0.5870 0.02906 0.05706
## 16 0.8046 0.6049 0.02248 0.04531
## 32 0.8118 0.6190 0.02555 0.05141
## 64 0.8183 0.6330 0.02513 0.05031
## 70 0.8258 0.6482 0.02137 0.04275 *
##
## The top 5 variables (out of 70):
## sprice.d20nexp, sprice.log10, sprice.root2, biddable, spdiff.cut.fctr(1,10]
##
## [1] "sprice.d20nexp"
## [2] "sprice.log10"
## [3] "sprice.root2"
## [4] "biddable"
## [5] "spdiff.cut.fctr(1,10]"
## [6] "spdiff.cut.fctr(-1,0]"
## [7] "spdiff.cut.fctr(10,100]"
## [8] "cellular.fctr0:carrier.fctrNone"
## [9] "spdiff.cut.fctr(-10,-1]"
## [10] "prdl.my.fctriPad1"
## [11] "spdiff.cut.fctr(0,1]"
## [12] "D.ratio.wrds.stop.n.wrds.n"
## [13] "storage.fctr16"
## [14] "condition.fctrFor parts or not working"
## [15] "D.weight.sum.stem.stop.Ratio"
## [16] "color.fctrUnknown"
## [17] "prdl.my.fctriPad3"
## [18] "startprice.dcm2.is9"
## [19] "prdl.my.fctriPad2"
## [20] ".rnorm"
## [21] "storage.fctr32"
## [22] "prdl.my.fctriPadmini"
## [23] "cellular.fctr1:carrier.fctrVerizon"
## [24] "D.chrs.pnct11.n.log"
## [25] "prdl.my.fctriPad1:.clusterid.fctr2"
## [26] "prdl.my.fctriPadAir:.clusterid.fctr2"
## [27] "prdl.my.fctriPadmini2"
## [28] "prdl.my.fctriPad1:.clusterid.fctr3"
## [29] "prdl.my.fctriPad4:.clusterid.fctr2"
## [30] "prdl.my.fctriPad3:.clusterid.fctr2"
## [31] "cellular.fctr1:carrier.fctrSprint"
## [32] "storage.fctr64"
## [33] "prdl.my.fctrUnknown:.clusterid.fctr3"
## [34] "D.wrds.stop.n.log"
## [35] "cellular.fctr1:carrier.fctrT-Mobile"
## [36] "prdl.my.fctrUnknown:.clusterid.fctr2"
## [37] "condition.fctrManufacturer refurbished"
## [38] "startprice.dcm1.is9"
## [39] "condition.fctrNew other (see details)"
## [40] "prdl.my.fctriPadAir"
## [41] "prdl.my.fctriPad4"
## [42] "color.fctrGold"
## [43] "cellular.fctr1:carrier.fctrUnknown"
## [44] "prdl.my.fctriPad4:.clusterid.fctr3"
## [45] "startprice.dgt2.is9"
## [46] "color.fctrWhite"
## [47] "prdl.my.fctriPadAir2"
## [48] "D.ratio.weight.sum.wrds.n"
## [49] "storage.fctrUnknown"
## [50] "color.fctrSpace Gray"
## [51] "spdiff.cut.fctr(100,1e+03]"
## [52] "D.chrs.pnct13.n.log"
## [53] "cellular.fctrUnknown"
## [54] "cellular.fctrUnknown:carrier.fctrUnknown"
## [55] "prdl.my.fctriPadmini3"
## [56] "cellular.fctr1"
## [57] "condition.fctrSeller refurbished"
## [58] "D.weight.post.stop.sum"
## [59] "D.weight.post.stem.sum"
## [60] "D.weight.sum"
## [61] "D.wrds.n.log"
## [62] "D.chrs.uppr.n.log"
## [63] "D.wrds.unq.n.log"
## [64] "D.terms.post.stop.n.log"
## [65] "D.terms.post.stem.n.log"
## [66] "D.chrs.n.log"
## [67] "condition.fctrNew"
## [68] "startprice.dgt1.is9"
## [69] "spdiff.cut.fctr(-100,-10]"
## [70] "spdiff.cut.fctr(-1e+03,-100]"
# print(all.equal(rfe_results[-which(names(rfe_results) == "times")],
# sav_rfe_results[-which(names(sav_rfe_results) == "times")]))
# require(mRMRe)
# indep_vars_vctr <- subset(glb_feats_df, !nzv &
# (exclude.as.feat != 1))[, "id"]
# indep_vars_vctr <- setdiff(indep_vars_vctr,
# myfind_fctr_cols_df(glb_trnobs_df[, c(glb_rsp_var, indep_vars_vctr)]))
# tmp_trnobs_df <- glb_trnobs_df[, c(glb_rsp_var, indep_vars_vctr)]
# tmp_trnobs_df$biddable <- as.numeric(tmp_trnobs_df$biddable)
# dd <- mRMR.data(data = tmp_trnobs_df)
# mRMRe.fltr <- mRMR.classic(data = dd, target_indices = c(1), feature_count = 10)
# print(solutions(mRMRe.fltr)[[1]])
# print(apply(solutions(mRMRe.fltr)[[1]], 2, function(x, y) { return(y[x]) },
# y=featureNames(dd)))
# print(featureNames(dd)[solutions(mRMRe.fltr)[[1]]])
# print(mRMRe.fltr@filters); print(mRMRe.fltr@scores)
mycheck_problem_data(glb_allobs_df, featsExclude = glbFeatsExclude,
fctrMaxUniqVals = glbFctrMaxUniqVals, terminate = TRUE)## [1] "numeric data missing in : "
## sold sold.fctr
## 798 798
## [1] "numeric data w/ 0s in : "
## biddable sold
## 1437 995
## sprice.log10 cellular.fctr
## 31 1590
## D.terms.post.stop.n D.terms.post.stop.n.log
## 1516 1516
## D.weight.post.stop.sum D.terms.post.stem.n
## 1516 1516
## D.terms.post.stem.n.log D.weight.post.stem.sum
## 1516 1516
## D.T.condit D.T.use
## 2156 2363
## D.T.in. D.T.good
## 2216 2451
## D.T.screen D.T.with
## 2440 2434
## D.T.of D.T.mint
## 2497 2585
## D.T.or D.T.cosmet
## 2523 2531
## D.T.minor D.T.light
## 2531 2567
## D.T.X100 D.T.from
## 2584 2591
## D.T.hous D.wrds.n.log
## 2576 1513
## D.wrds.unq.n.log D.weight.sum
## 1516 1516
## D.ratio.weight.sum.wrds.n D.chrs.n.log
## 1516 1513
## D.chrs.uppr.n.log D.dgts.n.log
## 1515 2452
## D.chrs.pnct01.n.log D.chrs.pnct02.n.log
## 2570 2648
## D.chrs.pnct03.n.log D.chrs.pnct04.n.log
## 2647 2648
## D.chrs.pnct05.n.log D.chrs.pnct06.n.log
## 2582 2596
## D.chrs.pnct07.n.log D.chrs.pnct08.n.log
## 2611 2572
## D.chrs.pnct09.n.log D.chrs.pnct10.n.log
## 2632 2639
## D.chrs.pnct11.n.log D.chrs.pnct12.n.log
## 2294 2534
## D.chrs.pnct13.n.log D.chrs.pnct14.n.log
## 1930 2574
## D.chrs.pnct15.n.log D.chrs.pnct16.n.log
## 2628 2639
## D.chrs.pnct17.n.log D.chrs.pnct18.n.log
## 2646 2647
## D.chrs.pnct19.n.log D.chrs.pnct20.n.log
## 2648 2648
## D.chrs.pnct21.n.log D.chrs.pnct22.n.log
## 2648 2648
## D.chrs.pnct23.n.log D.chrs.pnct24.n.log
## 2648 2648
## D.chrs.pnct25.n.log D.chrs.pnct26.n.log
## 2648 2648
## D.chrs.pnct27.n.log D.chrs.pnct28.n.log
## 2648 2640
## D.chrs.pnct29.n.log D.chrs.pnct30.n.log
## 2648 2648
## D.wrds.stop.n.log D.P.http
## 1965 2648
## D.P.mini D.P.air
## 2615 2627
## D.P.black D.P.white
## 2631 2638
## D.P.gold D.P.spacegray
## 2647 2642
## startprice.dgt1.is9 startprice.dgt2.is9
## 1779 2290
## startprice.dgt3.is9 startprice.dcm1.is9
## 2643 1653
## startprice.dcm2.is9
## 1826
## [1] "numeric data w/ Infs in : "
## named integer(0)
## [1] "numeric data w/ NaNs in : "
## named integer(0)
## [1] "string data missing in : "
## description condition cellular carrier color storage
## 1513 0 0 0 0 0
## productline .grpid descr.my .lcn
## 0 NA 1513 798# glb_allobs_df %>% filter(is.na(Married.fctr)) %>% tbl_df()
# glb_allobs_df %>% count(Married.fctr)
# levels(glb_allobs_df$Married.fctr)
print("glb_feats_df:"); print(dim(glb_feats_df))## [1] "glb_feats_df:"## [1] 93 12sav_feats_df <- glb_feats_df
glb_feats_df <- sav_feats_df
glb_feats_df[, "rsp_var_raw"] <- FALSE
glb_feats_df[glb_feats_df$id == glb_rsp_var_raw, "rsp_var_raw"] <- TRUE
glb_feats_df$exclude.as.feat <- (glb_feats_df$exclude.as.feat == 1)
if (!is.null(glb_id_var) && glb_id_var != ".rownames")
glb_feats_df[glb_feats_df$id %in% glb_id_var, "id_var"] <- TRUE
add_feats_df <- data.frame(id=glb_rsp_var, exclude.as.feat=TRUE, rsp_var=TRUE)
row.names(add_feats_df) <- add_feats_df$id; print(add_feats_df)## id exclude.as.feat rsp_var
## sold.fctr sold.fctr TRUE TRUEglb_feats_df <- myrbind_df(glb_feats_df, add_feats_df)
if (glb_id_var != ".rownames")
print(subset(glb_feats_df, rsp_var_raw | rsp_var | id_var)) else
print(subset(glb_feats_df, rsp_var_raw | rsp_var)) ## id cor.y exclude.as.feat cor.y.abs cor.high.X
## sold sold 1.0000000 TRUE 1.0000000 <NA>
## UniqueID UniqueID -0.1906261 TRUE 0.1906261 <NA>
## sold.fctr sold.fctr NA TRUE NA <NA>
## freqRatio percentUnique zeroVar nzv is.cor.y.abs.low
## sold 1.163743 0.1081081 FALSE FALSE FALSE
## UniqueID 1.000000 100.0000000 FALSE FALSE FALSE
## sold.fctr NA NA NA NA NA
## interaction.feat shapiro.test.p.value rsp_var_raw id_var rsp_var
## sold <NA> NA TRUE NA NA
## UniqueID <NA> NA FALSE TRUE NA
## sold.fctr <NA> NA NA NA TRUEprint("glb_feats_df vs. glb_allobs_df: "); ## [1] "glb_feats_df vs. glb_allobs_df: "print(setdiff(glb_feats_df$id, names(glb_allobs_df)))## character(0)print("glb_allobs_df vs. glb_feats_df: "); ## [1] "glb_allobs_df vs. glb_feats_df: "# Ensure these are only chr vars
print(setdiff(setdiff(names(glb_allobs_df), glb_feats_df$id),
myfind_chr_cols_df(glb_allobs_df)))## character(0)if (glb_save_envir)
save(glb_feats_df,
glb_allobs_df, #glb_trnobs_df, glb_fitobs_df, glb_OOBobs_df, glb_newobs_df,
file=paste0(glb_out_pfx, "selfts_dsk.RData"))
# load(paste0(glb_out_pfx, "blddfs_dsk.RData"))
# if (!all.equal(tmp_feats_df, glb_feats_df))
# stop("glb_feats_df r/w not working")
# if (!all.equal(tmp_entity_df, glb_allobs_df))
# stop("glb_allobs_df r/w not working")
glb_chunks_df <- myadd_chunk(glb_chunks_df, "fit.models", major.inc=TRUE)## label step_major step_minor label_minor bgn end
## 9 select.features 5 0 0 331.640 342.849
## 10 fit.models 6 0 0 342.849 NA
## elapsed
## 9 11.209
## 10 NAStep 6.0: fit models
# load(paste0(glb_out_pfx, "dsk.RData"))
get_model_sel_frmla <- function() {
model_evl_terms <- c(NULL)
# min.aic.fit might not be avl
lclMdlEvlCriteria <-
glbMdlMetricsEval[glbMdlMetricsEval %in% names(glb_models_df)]
for (metric in lclMdlEvlCriteria)
model_evl_terms <- c(model_evl_terms,
ifelse(length(grep("max", metric)) > 0, "-", "+"), metric)
if (glb_is_classification && glb_is_binomial)
model_evl_terms <- c(model_evl_terms, "-", "opt.prob.threshold.OOB")
model_sel_frmla <- as.formula(paste(c("~ ", model_evl_terms), collapse = " "))
return(model_sel_frmla)
}
get_dsp_models_df <- function() {
dsp_models_cols <- c("id",
glbMdlMetricsEval[glbMdlMetricsEval %in% names(glb_models_df)],
grep("opt.", names(glb_models_df), fixed = TRUE, value = TRUE))
dsp_models_df <-
#orderBy(get_model_sel_frmla(), glb_models_df)[, c("id", glbMdlMetricsEval)]
orderBy(get_model_sel_frmla(), glb_models_df)[, dsp_models_cols]
nCvMdl <- sapply(glb_models_lst, function(mdl) nrow(mdl$results))
nParams <- sapply(glb_models_lst, function(mdl) ifelse(mdl$method == "custom", 0,
nrow(subset(modelLookup(mdl$method), parameter != "parameter"))))
# nCvMdl <- nCvMdl[names(nCvMdl) != "avNNet"]
# nParams <- nParams[names(nParams) != "avNNet"]
if (length(cvMdlProblems <- nCvMdl[nCvMdl <= nParams]) > 0) {
print("Cross Validation issues:")
warning("Cross Validation issues:")
print(cvMdlProblems)
}
pltMdls <- setdiff(names(nCvMdl), names(cvMdlProblems))
pltMdls <- setdiff(pltMdls, names(nParams[nParams == 0]))
# length(pltMdls) == 21
png(paste0(glb_out_pfx, "bestTune.png"), width = 480 * 2, height = 480 * 4)
grid.newpage()
pushViewport(viewport(layout = grid.layout(ceiling(length(pltMdls) / 2.0), 2)))
pltIx <- 1
for (mdlId in pltMdls) {
print(ggplot(glb_models_lst[[mdlId]], highBestTune = TRUE) + labs(title = mdlId),
vp = viewport(layout.pos.row = ceiling(pltIx / 2.0),
layout.pos.col = ((pltIx - 1) %% 2) + 1))
pltIx <- pltIx + 1
}
dev.off()
return(dsp_models_df)
}
#get_dsp_models_df()
if (glb_is_classification && glb_is_binomial &&
(length(unique(glb_fitobs_df[, glb_rsp_var])) < 2))
stop("glb_fitobs_df$", glb_rsp_var, ": contains less than 2 unique values: ",
paste0(unique(glb_fitobs_df[, glb_rsp_var]), collapse=", "))
max_cor_y_x_vars <- orderBy(~ -cor.y.abs,
subset(glb_feats_df, (exclude.as.feat == 0) & !nzv & !is.cor.y.abs.low &
is.na(cor.high.X)))[1:2, "id"]
# while(length(max_cor_y_x_vars) < 2) {
# max_cor_y_x_vars <- c(max_cor_y_x_vars, orderBy(~ -cor.y.abs,
# subset(glb_feats_df, (exclude.as.feat == 0) & !is.cor.y.abs.low))[3, "id"])
# }
#stop(here"); glb_to_sav(); glb_models_lst <- sav_models_lst; glb_models_df <- sav_models_df
if (!is.null(glb_Baseline_mdl_var)) {
if ((max_cor_y_x_vars[1] != glb_Baseline_mdl_var) &
(glb_feats_df[glb_feats_df$id == max_cor_y_x_vars[1], "cor.y.abs"] >
glb_feats_df[glb_feats_df$id == glb_Baseline_mdl_var, "cor.y.abs"]))
stop(max_cor_y_x_vars[1], " has a higher correlation with ", glb_rsp_var,
" than the Baseline var: ", glb_Baseline_mdl_var)
}
glb_model_type <- ifelse(glb_is_regression, "regression", "classification")
# Model specs
c("id.prefix", "method", "type",
# trainControl params
"preProc.method", "cv.n.folds", "cv.n.repeats", "summary.fn",
# train params
"metric", "metric.maximize", "tune.df")## [1] "id.prefix" "method" "type"
## [4] "preProc.method" "cv.n.folds" "cv.n.repeats"
## [7] "summary.fn" "metric" "metric.maximize"
## [10] "tune.df"# Baseline
if (!is.null(glb_Baseline_mdl_var))
ret_lst <- myfit_mdl(mdl_id="Baseline",
model_method="mybaseln_classfr",
indep_vars_vctr=glb_Baseline_mdl_var,
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_fitobs_df, OOB_df=glb_OOBobs_df)
# Most Frequent Outcome "MFO" model: mean(y) for regression
# Not using caret's nullModel since model stats not avl
# Cannot use rpart for multinomial classification since it predicts non-MFO
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "MFO", type = glb_model_type, trainControl.method = "none",
train.method = ifelse(glb_is_regression, "lm", "myMFO_classfr"))),
indep_vars = ".rnorm", rsp_var = glb_rsp_var,
fit_df = glb_fitobs_df, OOB_df = glb_OOBobs_df)## [1] "fitting model: MFO.myMFO_classfr"
## [1] " indep_vars: .rnorm"## Warning in if (mdl_specs_lst[["train.method"]] %in% c("bayesglm", "glm", :
## the condition has length > 1 and only the first element will be used## Fitting parameter = none on full training set
## [1] "in MFO.Classifier$fit"
## [1] "unique.vals:"
## [1] N Y
## Levels: N Y
## [1] "unique.prob:"
## y
## N Y
## 0.5372549 0.4627451
## [1] "MFO.val:"
## [1] "N"
## Length Class Mode
## unique.vals 2 factor numeric
## unique.prob 2 -none- numeric
## MFO.val 1 -none- character
## x.names 1 -none- character
## xNames 1 -none- character
## problemType 1 -none- character
## tuneValue 1 data.frame list
## obsLevels 2 -none- character
## [1] "entr MFO.Classifier$predict"
## [1] "exit MFO.Classifier$predict"## Loading required package: ROCR
## Loading required package: gplots
##
## Attaching package: 'gplots'
##
## The following object is masked from 'package:wordcloud':
##
## textplot
##
## The following object is masked from 'package:stats':
##
## lowess## [1] "in MFO.Classifier$prob"
## N Y
## 1 0.5372549 0.4627451
## 2 0.5372549 0.4627451
## 3 0.5372549 0.4627451
## 4 0.5372549 0.4627451
## 5 0.5372549 0.4627451
## 6 0.5372549 0.4627451
## sold.fctr sold.fctr.predict.MFO.myMFO_classfr.Y
## 1 N 548
## 2 Y 472
## Prediction
## Reference N Y
## N 0 548
## Y 0 472
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 4.627451e-01 0.000000e+00 4.318011e-01 4.939047e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 9.999992e-01 9.327417e-121
## [1] "entr MFO.Classifier$predict"
## [1] "exit MFO.Classifier$predict"
## [1] "in MFO.Classifier$prob"
## N Y
## 1 0.5372549 0.4627451
## 2 0.5372549 0.4627451
## 3 0.5372549 0.4627451
## 4 0.5372549 0.4627451
## 5 0.5372549 0.4627451
## 6 0.5372549 0.4627451
## sold.fctr sold.fctr.predict.MFO.myMFO_classfr.Y
## 1 N 447
## 2 Y 383
## Prediction
## Reference N Y
## N 0 447
## Y 0 383
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 4.614458e-01 0.000000e+00 4.271177e-01 4.960486e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 9.999963e-01 8.826336e-99
## id feats max.nTuningRuns min.elapsedtime.everything
## 1 MFO.myMFO_classfr .rnorm 0 0.292
## min.elapsedtime.final max.AUCpROC.fit max.Sens.fit max.Spec.fit
## 1 0.003 0.5 1 0
## max.AUCROCR.fit opt.prob.threshold.fit max.f.score.fit max.Accuracy.fit
## 1 0.5 0.4 0.6327078 0.4627451
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.4318011 0.4939047 0
## max.AUCpROC.OOB max.Sens.OOB max.Spec.OOB max.AUCROCR.OOB
## 1 0.5 1 0 0.5
## opt.prob.threshold.OOB max.f.score.OOB max.Accuracy.OOB
## 1 0.4 0.6314922 0.4614458
## max.AccuracyLower.OOB max.AccuracyUpper.OOB max.Kappa.OOB
## 1 0.4271177 0.4960486 0if (glb_is_classification)
# "random" model - only for classification;
# none needed for regression since it is same as MFO
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = "Random", type = glb_model_type, trainControl.method = "none",
train.method = "myrandom_classfr")),
indep_vars = ".rnorm", rsp_var = glb_rsp_var,
fit_df = glb_fitobs_df, OOB_df = glb_OOBobs_df)## [1] "fitting model: Random.myrandom_classfr"
## [1] " indep_vars: .rnorm"## Warning in if (mdl_specs_lst[["train.method"]] %in% c("bayesglm", "glm", :
## the condition has length > 1 and only the first element will be used## Fitting parameter = none on full training set
## Length Class Mode
## unique.vals 2 factor numeric
## unique.prob 2 table numeric
## xNames 1 -none- character
## problemType 1 -none- character
## tuneValue 1 data.frame list
## obsLevels 2 -none- character
## [1] "in Random.Classifier$prob"
## sold.fctr sold.fctr.predict.Random.myrandom_classfr.Y
## 1 N 548
## 2 Y 472
## Prediction
## Reference N Y
## N 0 548
## Y 0 472
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 4.627451e-01 0.000000e+00 4.318011e-01 4.939047e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 9.999992e-01 9.327417e-121
## [1] "in Random.Classifier$prob"
## sold.fctr sold.fctr.predict.Random.myrandom_classfr.Y
## 1 N 447
## 2 Y 383
## Prediction
## Reference N Y
## N 0 447
## Y 0 383
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 4.614458e-01 0.000000e+00 4.271177e-01 4.960486e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 9.999963e-01 8.826336e-99
## id feats max.nTuningRuns
## 1 Random.myrandom_classfr .rnorm 0
## min.elapsedtime.everything min.elapsedtime.final max.AUCpROC.fit
## 1 0.308 0.001 0.4859659
## max.Sens.fit max.Spec.fit max.AUCROCR.fit opt.prob.threshold.fit
## 1 0.5164234 0.4555085 0.4893217 0.4
## max.f.score.fit max.Accuracy.fit max.AccuracyLower.fit
## 1 0.6327078 0.4627451 0.4318011
## max.AccuracyUpper.fit max.Kappa.fit max.AUCpROC.OOB max.Sens.OOB
## 1 0.4939047 0 0.5064252 0.5480984
## max.Spec.OOB max.AUCROCR.OOB opt.prob.threshold.OOB max.f.score.OOB
## 1 0.464752 0.4956046 0.4 0.6314922
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.4614458 0.4271177 0.4960486
## max.Kappa.OOB
## 1 0# ret_lst <- myfit_mdl(mdl_id = "Random", model_method = "myrandom_classfr",
# model_type = glb_model_type,
# indep_vars_vctr = ".rnorm",
# rsp_var = glb_rsp_var, rsp_var_out = glb_rsp_var_out,
# fit_df = glb_fitobs_df, OOB_df = glb_OOBobs_df)
# Any models that have tuning parameters has "better" results with cross-validation
# (except bag & rf) & "different" results for different outcome metrics
# Max.cor.Y
# Check impact of cv
# rpart is not a good candidate since caret does not optimize cp (only tuning parameter of rpart) well
ret_lst <- myfit_mdl(mdl_specs_lst=myinit_mdl_specs_lst(mdl_specs_lst=list(
id.prefix="Max.cor.Y.rcv.1X1", type=glb_model_type, trainControl.method="none",
train.method="glmnet")),
indep_vars=max_cor_y_x_vars, rsp_var=glb_rsp_var,
fit_df=glb_fitobs_df, OOB_df=glb_OOBobs_df)## [1] "fitting model: Max.cor.Y.rcv.1X1.glmnet"
## [1] " indep_vars: biddable,sprice.root2"## Loading required package: glmnet
## Loading required package: Matrix
##
## Attaching package: 'Matrix'
##
## The following object is masked from 'package:tidyr':
##
## expand
##
## Loaded glmnet 2.0-2## Fitting alpha = 0.1, lambda = 0.00576 on full training set
## Length Class Mode
## a0 79 -none- numeric
## beta 158 dgCMatrix S4
## df 79 -none- numeric
## dim 2 -none- numeric
## lambda 79 -none- numeric
## dev.ratio 79 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 2 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept) biddable sprice.root2
## 1.0277229 1.8660218 -0.1575942
## [1] "max lambda < lambdaOpt:"
## (Intercept) biddable sprice.root2
## 1.0358390 1.8708816 -0.1583838
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.1X1.glmnet.N
## 1 N 432
## 2 Y 104
## sold.fctr.predict.Max.cor.Y.rcv.1X1.glmnet.Y
## 1 116
## 2 368
## Prediction
## Reference N Y
## N 432 116
## Y 104 368
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.843137e-01 5.669826e-01 7.577802e-01 8.091950e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 1.365269e-60 4.583177e-01
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.1X1.glmnet.N
## 1 N 352
## 2 Y 94
## sold.fctr.predict.Max.cor.Y.rcv.1X1.glmnet.Y
## 1 95
## 2 289
## Prediction
## Reference N Y
## N 352 95
## Y 94 289
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.722892e-01 5.419399e-01 7.422177e-01 8.004125e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 2.113571e-44 1.000000e+00
## id feats max.nTuningRuns
## 1 Max.cor.Y.rcv.1X1.glmnet biddable,sprice.root2 0
## min.elapsedtime.everything min.elapsedtime.final max.AUCpROC.fit
## 1 0.74 0.023 0.7898522
## max.Sens.fit max.Spec.fit max.AUCROCR.fit opt.prob.threshold.fit
## 1 0.8339416 0.7457627 0.86206 0.4
## max.f.score.fit max.Accuracy.fit max.AccuracyLower.fit
## 1 0.7698745 0.7843137 0.7577802
## max.AccuracyUpper.fit max.Kappa.fit max.AUCpROC.OOB max.Sens.OOB
## 1 0.809195 0.5669826 0.7665376 0.8255034
## max.Spec.OOB max.AUCROCR.OOB opt.prob.threshold.OOB max.f.score.OOB
## 1 0.7075718 0.8197236 0.4 0.7535854
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.7722892 0.7422177 0.8004125
## max.Kappa.OOB
## 1 0.5419399# rcv_n_folds == 1 & rcv_n_repeats > 1 crashes
for (rcv_n_folds in seq(3, glb_rcv_n_folds + 2, 2))
for (rcv_n_repeats in seq(1, glb_rcv_n_repeats + 2, 2)) {
ret_lst <- myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = paste0("Max.cor.Y.rcv.", rcv_n_folds, "X", rcv_n_repeats),
type = glb_model_type, trainControl.method = "repeatedcv",
trainControl.number = rcv_n_folds, trainControl.repeats = rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.method = "glmnet", train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize)),
indep_vars = max_cor_y_x_vars, rsp_var = glb_rsp_var,
fit_df = glb_fitobs_df, OOB_df = glb_OOBobs_df)
}## [1] "fitting model: Max.cor.Y.rcv.3X1.glmnet"
## [1] " indep_vars: biddable,sprice.root2"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0.1, lambda = 0.124 on full training set## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst =
## list(id.prefix = paste0("Max.cor.Y.rcv.", : model's bestTune found at an
## extreme of tuneGrid for parameter: alpha## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst =
## list(id.prefix = paste0("Max.cor.Y.rcv.", : model's bestTune found at an
## extreme of tuneGrid for parameter: lambda
## Length Class Mode
## a0 79 -none- numeric
## beta 158 dgCMatrix S4
## df 79 -none- numeric
## dim 2 -none- numeric
## lambda 79 -none- numeric
## dev.ratio 79 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 2 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.39977062 1.21150458 -0.08600408
## [1] "max lambda < lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.4231996 1.2510219 -0.0892442
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.3X1.glmnet.N
## 1 N 457
## 2 Y 118
## sold.fctr.predict.Max.cor.Y.rcv.3X1.glmnet.Y
## 1 91
## 2 354
## Prediction
## Reference N Y
## N 457 91
## Y 118 354
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.950980e-01 5.862671e-01 7.690020e-01 8.194783e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 3.331019e-66 7.210452e-02
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.3X1.glmnet.N
## 1 N 342
## 2 Y 89
## sold.fctr.predict.Max.cor.Y.rcv.3X1.glmnet.Y
## 1 105
## 2 294
## Prediction
## Reference N Y
## N 342 105
## Y 89 294
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.662651e-01 5.311363e-01 7.359543e-01 7.946712e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 4.058545e-42 2.815083e-01
## id feats max.nTuningRuns
## 1 Max.cor.Y.rcv.3X1.glmnet biddable,sprice.root2 25
## min.elapsedtime.everything min.elapsedtime.final max.AUCpROC.fit
## 1 1.546 0.018 0.7919708
## max.Sens.fit max.Spec.fit max.AUCROCR.fit opt.prob.threshold.fit
## 1 0.8339416 0.75 0.8621335 0.5
## max.f.score.fit max.Accuracy.fit max.AccuracyLower.fit
## 1 0.7720829 0.7950911 0.769002
## max.AccuracyUpper.fit max.Kappa.fit max.AUCpROC.OOB max.Sens.OOB
## 1 0.8194783 0.5860164 0.7592654 0.8187919
## max.Spec.OOB max.AUCROCR.OOB opt.prob.threshold.OOB max.f.score.OOB
## 1 0.6997389 0.8197353 0.4 0.7519182
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.7662651 0.7359543 0.7946712
## max.Kappa.OOB max.AccuracySD.fit max.KappaSD.fit
## 1 0.5311363 0.003934893 0.009324777
## [1] "fitting model: Max.cor.Y.rcv.3X3.glmnet"
## [1] " indep_vars: biddable,sprice.root2"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0.325, lambda = 0.124 on full training set## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst =
## list(id.prefix = paste0("Max.cor.Y.rcv.", : model's bestTune found at an
## extreme of tuneGrid for parameter: lambda
## Length Class Mode
## a0 67 -none- numeric
## beta 134 dgCMatrix S4
## df 67 -none- numeric
## dim 2 -none- numeric
## lambda 67 -none- numeric
## dev.ratio 67 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 2 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.36344652 1.18295779 -0.08216503
## [1] "max lambda < lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.39118688 1.22752496 -0.08592317
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.3X3.glmnet.N
## 1 N 457
## 2 Y 118
## sold.fctr.predict.Max.cor.Y.rcv.3X3.glmnet.Y
## 1 91
## 2 354
## Prediction
## Reference N Y
## N 457 91
## Y 118 354
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.950980e-01 5.862671e-01 7.690020e-01 8.194783e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 3.331019e-66 7.210452e-02
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.3X3.glmnet.N
## 1 N 341
## 2 Y 88
## sold.fctr.predict.Max.cor.Y.rcv.3X3.glmnet.Y
## 1 106
## 2 295
## Prediction
## Reference N Y
## N 341 106
## Y 88 295
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.662651e-01 5.313109e-01 7.359543e-01 7.946712e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 4.058545e-42 2.222645e-01
## id feats max.nTuningRuns
## 1 Max.cor.Y.rcv.3X3.glmnet biddable,sprice.root2 25
## min.elapsedtime.everything min.elapsedtime.final max.AUCpROC.fit
## 1 2.078 0.016 0.7919708
## max.Sens.fit max.Spec.fit max.AUCROCR.fit opt.prob.threshold.fit
## 1 0.8339416 0.75 0.8622147 0.5
## max.f.score.fit max.Accuracy.fit max.AccuracyLower.fit
## 1 0.7720829 0.7954221 0.769002
## max.AccuracyUpper.fit max.Kappa.fit max.AUCpROC.OOB max.Sens.OOB
## 1 0.8194783 0.5866808 0.760384 0.8210291
## max.Spec.OOB max.AUCROCR.OOB opt.prob.threshold.OOB max.f.score.OOB
## 1 0.6997389 0.819747 0.4 0.752551
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.7662651 0.7359543 0.7946712
## max.Kappa.OOB max.AccuracySD.fit max.KappaSD.fit
## 1 0.5313109 0.0200887 0.04087803
## [1] "fitting model: Max.cor.Y.rcv.3X5.glmnet"
## [1] " indep_vars: biddable,sprice.root2"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0.55, lambda = 0.124 on full training set## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst =
## list(id.prefix = paste0("Max.cor.Y.rcv.", : model's bestTune found at an
## extreme of tuneGrid for parameter: lambda
## Length Class Mode
## a0 61 -none- numeric
## beta 122 dgCMatrix S4
## df 61 -none- numeric
## dim 2 -none- numeric
## lambda 61 -none- numeric
## dev.ratio 61 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 2 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.28887900 1.10455877 -0.07355771
## [1] "max lambda < lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.32191930 1.15672712 -0.07800633
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.3X5.glmnet.N
## 1 N 457
## 2 Y 118
## sold.fctr.predict.Max.cor.Y.rcv.3X5.glmnet.Y
## 1 91
## 2 354
## Prediction
## Reference N Y
## N 457 91
## Y 118 354
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.950980e-01 5.862671e-01 7.690020e-01 8.194783e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 3.331019e-66 7.210452e-02
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.3X5.glmnet.N
## 1 N 340
## 2 Y 88
## sold.fctr.predict.Max.cor.Y.rcv.3X5.glmnet.Y
## 1 107
## 2 295
## Prediction
## Reference N Y
## N 340 107
## Y 88 295
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.650602e-01 5.289828e-01 7.347026e-01 7.935220e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 1.138374e-41 1.973957e-01
## id feats max.nTuningRuns
## 1 Max.cor.Y.rcv.3X5.glmnet biddable,sprice.root2 25
## min.elapsedtime.everything min.elapsedtime.final max.AUCpROC.fit
## 1 2.568 0.016 0.7919708
## max.Sens.fit max.Spec.fit max.AUCROCR.fit opt.prob.threshold.fit
## 1 0.8339416 0.75 0.8622069 0.5
## max.f.score.fit max.Accuracy.fit max.AccuracyLower.fit
## 1 0.7720829 0.7962655 0.769002
## max.AccuracyUpper.fit max.Kappa.fit max.AUCpROC.OOB max.Sens.OOB
## 1 0.8194783 0.58834 0.760384 0.8210291
## max.Spec.OOB max.AUCROCR.OOB opt.prob.threshold.OOB max.f.score.OOB
## 1 0.6997389 0.8196593 0.4 0.7515924
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.7650602 0.7347026 0.793522
## max.Kappa.OOB max.AccuracySD.fit max.KappaSD.fit
## 1 0.5289828 0.01812066 0.03689711
## [1] "fitting model: Max.cor.Y.rcv.5X1.glmnet"
## [1] " indep_vars: biddable,sprice.root2"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0.775, lambda = 0.124 on full training set## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst =
## list(id.prefix = paste0("Max.cor.Y.rcv.", : model's bestTune found at an
## extreme of tuneGrid for parameter: lambda
## Length Class Mode
## a0 57 -none- numeric
## beta 114 dgCMatrix S4
## df 57 -none- numeric
## dim 2 -none- numeric
## lambda 57 -none- numeric
## dev.ratio 57 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 2 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.19570698 1.01735782 -0.06318327
## [1] "max lambda < lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.23591040 1.07863496 -0.06852179
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.5X1.glmnet.N
## 1 N 461
## 2 Y 118
## sold.fctr.predict.Max.cor.Y.rcv.5X1.glmnet.Y
## 1 87
## 2 354
## Prediction
## Reference N Y
## N 461 87
## Y 118 354
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.990196e-01 5.939459e-01 7.730893e-01 8.232111e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 2.536091e-68 3.614514e-02
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.5X1.glmnet.N
## 1 N 339
## 2 Y 88
## sold.fctr.predict.Max.cor.Y.rcv.5X1.glmnet.Y
## 1 108
## 2 295
## Prediction
## Reference N Y
## N 339 108
## Y 88 295
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.638554e-01 5.266555e-01 7.334512e-01 7.923725e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 3.171799e-41 1.747358e-01
## id feats max.nTuningRuns
## 1 Max.cor.Y.rcv.5X1.glmnet biddable,sprice.root2 25
## min.elapsedtime.everything min.elapsedtime.final max.AUCpROC.fit
## 1 1.729 0.015 0.7956204
## max.Sens.fit max.Spec.fit max.AUCROCR.fit opt.prob.threshold.fit
## 1 0.8412409 0.75 0.8621064 0.5
## max.f.score.fit max.Accuracy.fit max.AccuracyLower.fit
## 1 0.7754655 0.7990899 0.7730893
## max.AccuracyUpper.fit max.Kappa.fit max.AUCpROC.OOB max.Sens.OOB
## 1 0.8232111 0.5944245 0.7615026 0.8232662
## max.Spec.OOB max.AUCROCR.OOB opt.prob.threshold.OOB max.f.score.OOB
## 1 0.6997389 0.81956 0.4 0.7506361
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.7638554 0.7334512 0.7923725
## max.Kappa.OOB max.AccuracySD.fit max.KappaSD.fit
## 1 0.5266555 0.03608644 0.07068726
## [1] "fitting model: Max.cor.Y.rcv.5X3.glmnet"
## [1] " indep_vars: biddable,sprice.root2"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0.775, lambda = 0.124 on full training set## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst =
## list(id.prefix = paste0("Max.cor.Y.rcv.", : model's bestTune found at an
## extreme of tuneGrid for parameter: lambda
## Length Class Mode
## a0 57 -none- numeric
## beta 114 dgCMatrix S4
## df 57 -none- numeric
## dim 2 -none- numeric
## lambda 57 -none- numeric
## dev.ratio 57 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 2 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.19570698 1.01735782 -0.06318327
## [1] "max lambda < lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.23591040 1.07863496 -0.06852179
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.5X3.glmnet.N
## 1 N 461
## 2 Y 118
## sold.fctr.predict.Max.cor.Y.rcv.5X3.glmnet.Y
## 1 87
## 2 354
## Prediction
## Reference N Y
## N 461 87
## Y 118 354
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.990196e-01 5.939459e-01 7.730893e-01 8.232111e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 2.536091e-68 3.614514e-02
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.5X3.glmnet.N
## 1 N 339
## 2 Y 88
## sold.fctr.predict.Max.cor.Y.rcv.5X3.glmnet.Y
## 1 108
## 2 295
## Prediction
## Reference N Y
## N 339 108
## Y 88 295
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.638554e-01 5.266555e-01 7.334512e-01 7.923725e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 3.171799e-41 1.747358e-01
## id feats max.nTuningRuns
## 1 Max.cor.Y.rcv.5X3.glmnet biddable,sprice.root2 25
## min.elapsedtime.everything min.elapsedtime.final max.AUCpROC.fit
## 1 2.597 0.016 0.7956204
## max.Sens.fit max.Spec.fit max.AUCROCR.fit opt.prob.threshold.fit
## 1 0.8412409 0.75 0.8621064 0.5
## max.f.score.fit max.Accuracy.fit max.AccuracyLower.fit
## 1 0.7754655 0.7964364 0.7730893
## max.AccuracyUpper.fit max.Kappa.fit max.AUCpROC.OOB max.Sens.OOB
## 1 0.8232111 0.5888026 0.7615026 0.8232662
## max.Spec.OOB max.AUCROCR.OOB opt.prob.threshold.OOB max.f.score.OOB
## 1 0.6997389 0.81956 0.4 0.7506361
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.7638554 0.7334512 0.7923725
## max.Kappa.OOB max.AccuracySD.fit max.KappaSD.fit
## 1 0.5266555 0.03349187 0.06729278
## [1] "fitting model: Max.cor.Y.rcv.5X5.glmnet"
## [1] " indep_vars: biddable,sprice.root2"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 1, lambda = 0.124 on full training set## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst =
## list(id.prefix = paste0("Max.cor.Y.rcv.", : model's bestTune found at an
## extreme of tuneGrid for parameter: alpha## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst =
## list(id.prefix = paste0("Max.cor.Y.rcv.", : model's bestTune found at an
## extreme of tuneGrid for parameter: lambda
## Length Class Mode
## a0 54 -none- numeric
## beta 108 dgCMatrix S4
## df 54 -none- numeric
## dim 2 -none- numeric
## lambda 54 -none- numeric
## dev.ratio 54 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 2 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.12532879 0.99816406 -0.05704275
## [1] "max lambda < lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.17511926 1.06475879 -0.06331821
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.5X5.glmnet.N
## 1 N 463
## 2 Y 119
## sold.fctr.predict.Max.cor.Y.rcv.5X5.glmnet.Y
## 1 85
## 2 353
## Prediction
## Reference N Y
## N 463 85
## Y 119 353
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 8.000000e-01 5.957477e-01 7.741117e-01 8.241437e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 7.378773e-69 2.086258e-02
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.5X5.glmnet.N
## 1 N 338
## 2 Y 87
## sold.fctr.predict.Max.cor.Y.rcv.5X5.glmnet.Y
## 1 109
## 2 296
## Prediction
## Reference N Y
## N 338 109
## Y 87 296
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.638554e-01 5.268317e-01 7.334512e-01 7.923725e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 3.171799e-41 1.336144e-01
## id feats max.nTuningRuns
## 1 Max.cor.Y.rcv.5X5.glmnet biddable,sprice.root2 25
## min.elapsedtime.everything min.elapsedtime.final max.AUCpROC.fit
## 1 3.111 0.015 0.7963859
## max.Sens.fit max.Spec.fit max.AUCROCR.fit opt.prob.threshold.fit
## 1 0.8448905 0.7478814 0.8621064 0.5
## max.f.score.fit max.Accuracy.fit max.AccuracyLower.fit
## 1 0.7758242 0.7970687 0.7741117
## max.AccuracyUpper.fit max.Kappa.fit max.AUCpROC.OOB max.Sens.OOB
## 1 0.8241437 0.5901123 0.7641135 0.8232662
## max.Spec.OOB max.AUCROCR.OOB opt.prob.threshold.OOB max.f.score.OOB
## 1 0.7049608 0.819379 0.4 0.751269
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.7638554 0.7334512 0.7923725
## max.Kappa.OOB max.AccuracySD.fit max.KappaSD.fit
## 1 0.5268317 0.03129087 0.06285747# Add parallel coordinates graph of glb_models_df[, glbMdlMetricsEval] to evaluate cv parameters
tmp_models_cols <- c("id", "max.nTuningRuns",
glbMdlMetricsEval[glbMdlMetricsEval %in% names(glb_models_df)],
grep("opt.", names(glb_models_df), fixed = TRUE, value = TRUE))
print(myplot_parcoord(obs_df = subset(glb_models_df,
grepl("Max.cor.Y.rcv.", id, fixed = TRUE),
select = -feats)[, tmp_models_cols],
id_var = "id"))
ret_lst <- myfit_mdl(mdl_specs_lst=myinit_mdl_specs_lst(mdl_specs_lst=list(
id.prefix="Max.cor.Y.rcv.1X1.cp.0", type=glb_model_type, trainControl.method="none",
train.method="rpart",
tune.df=data.frame(method="rpart", parameter="cp", min=0.0, max=0.0, by=0.1))),
indep_vars=max_cor_y_x_vars, rsp_var=glb_rsp_var,
fit_df=glb_fitobs_df, OOB_df=glb_OOBobs_df)## [1] "fitting model: Max.cor.Y.rcv.1X1.cp.0.rpart"
## [1] " indep_vars: biddable,sprice.root2"## Loading required package: rpart## Fitting cp = 0 on full training set## Loading required package: rpart.plot
## Call:
## rpart(formula = .outcome ~ ., control = list(minsplit = 20, minbucket = 7,
## cp = 0, maxcompete = 4, maxsurrogate = 5, usesurrogate = 2,
## surrogatestyle = 0, maxdepth = 30, xval = 0))
## n= 1020
##
## CP nsplit rel error
## 1 0.546610169 0 1.0000000
## 2 0.025423729 1 0.4533898
## 3 0.010593220 2 0.4279661
## 4 0.006355932 5 0.3961864
## 5 0.003531073 6 0.3898305
## 6 0.002118644 9 0.3792373
## 7 0.001059322 16 0.3644068
## 8 0.000000000 20 0.3601695
##
## Variable importance
## biddable sprice.root2
## 52 48
##
## Node number 1: 1020 observations, complexity param=0.5466102
## predicted class=N expected loss=0.4627451 P(node) =1
## class counts: 548 472
## probabilities: 0.537 0.463
## left son=2 (554 obs) right son=3 (466 obs)
## Primary splits:
## biddable < 0.5 to the left, improve=169.2715, (0 missing)
## sprice.root2 < 10.23963 to the right, improve=133.6444, (0 missing)
## Surrogate splits:
## sprice.root2 < 12.26518 to the right, agree=0.751, adj=0.455, (0 split)
##
## Node number 2: 554 observations, complexity param=0.003531073
## predicted class=N expected loss=0.198556 P(node) =0.5431373
## class counts: 444 110
## probabilities: 0.801 0.199
## left son=4 (124 obs) right son=5 (430 obs)
## Primary splits:
## sprice.root2 < 19.90914 to the right, improve=3.855349, (0 missing)
##
## Node number 3: 466 observations, complexity param=0.02542373
## predicted class=Y expected loss=0.223176 P(node) =0.4568627
## class counts: 104 362
## probabilities: 0.223 0.777
## left son=6 (164 obs) right son=7 (302 obs)
## Primary splits:
## sprice.root2 < 11.83195 to the right, improve=49.71379, (0 missing)
##
## Node number 4: 124 observations
## predicted class=N expected loss=0.08870968 P(node) =0.1215686
## class counts: 113 11
## probabilities: 0.911 0.089
##
## Node number 5: 430 observations, complexity param=0.003531073
## predicted class=N expected loss=0.2302326 P(node) =0.4215686
## class counts: 331 99
## probabilities: 0.770 0.230
## left son=10 (263 obs) right son=11 (167 obs)
## Primary splits:
## sprice.root2 < 13.61047 to the right, improve=1.786236, (0 missing)
##
## Node number 6: 164 observations, complexity param=0.01059322
## predicted class=N expected loss=0.4634146 P(node) =0.1607843
## class counts: 88 76
## probabilities: 0.537 0.463
## left son=12 (19 obs) right son=13 (145 obs)
## Primary splits:
## sprice.root2 < 20.33187 to the right, improve=2.748634, (0 missing)
##
## Node number 7: 302 observations
## predicted class=Y expected loss=0.05298013 P(node) =0.2960784
## class counts: 16 286
## probabilities: 0.053 0.947
##
## Node number 10: 263 observations
## predicted class=N expected loss=0.1939163 P(node) =0.2578431
## class counts: 212 51
## probabilities: 0.806 0.194
##
## Node number 11: 167 observations, complexity param=0.003531073
## predicted class=N expected loss=0.2874251 P(node) =0.1637255
## class counts: 119 48
## probabilities: 0.713 0.287
## left son=22 (158 obs) right son=23 (9 obs)
## Primary splits:
## sprice.root2 < 13.41548 to the left, improve=4.574556, (0 missing)
##
## Node number 12: 19 observations
## predicted class=N expected loss=0.2105263 P(node) =0.01862745
## class counts: 15 4
## probabilities: 0.789 0.211
##
## Node number 13: 145 observations, complexity param=0.01059322
## predicted class=N expected loss=0.4965517 P(node) =0.1421569
## class counts: 73 72
## probabilities: 0.503 0.497
## left son=26 (87 obs) right son=27 (58 obs)
## Primary splits:
## sprice.root2 < 13.89077 to the right, improve=1.554023, (0 missing)
##
## Node number 22: 158 observations, complexity param=0.002118644
## predicted class=N expected loss=0.2594937 P(node) =0.154902
## class counts: 117 41
## probabilities: 0.741 0.259
## left son=44 (30 obs) right son=45 (128 obs)
## Primary splits:
## sprice.root2 < 12.54832 to the right, improve=1.884019, (0 missing)
##
## Node number 23: 9 observations
## predicted class=Y expected loss=0.2222222 P(node) =0.008823529
## class counts: 2 7
## probabilities: 0.222 0.778
##
## Node number 26: 87 observations, complexity param=0.01059322
## predicted class=N expected loss=0.4367816 P(node) =0.08529412
## class counts: 49 38
## probabilities: 0.563 0.437
## left son=52 (48 obs) right son=53 (39 obs)
## Primary splits:
## sprice.root2 < 16.50717 to the left, improve=2.291777, (0 missing)
##
## Node number 27: 58 observations, complexity param=0.006355932
## predicted class=Y expected loss=0.4137931 P(node) =0.05686275
## class counts: 24 34
## probabilities: 0.414 0.586
## left son=54 (9 obs) right son=55 (49 obs)
## Primary splits:
## sprice.root2 < 12.24684 to the left, improve=1.362421, (0 missing)
##
## Node number 44: 30 observations
## predicted class=N expected loss=0.1 P(node) =0.02941176
## class counts: 27 3
## probabilities: 0.900 0.100
##
## Node number 45: 128 observations, complexity param=0.002118644
## predicted class=N expected loss=0.296875 P(node) =0.1254902
## class counts: 90 38
## probabilities: 0.703 0.297
## left son=90 (95 obs) right son=91 (33 obs)
## Primary splits:
## sprice.root2 < 11.40153 to the left, improve=2.210706, (0 missing)
##
## Node number 52: 48 observations, complexity param=0.001059322
## predicted class=N expected loss=0.3333333 P(node) =0.04705882
## class counts: 32 16
## probabilities: 0.667 0.333
## left son=104 (7 obs) right son=105 (41 obs)
## Primary splits:
## sprice.root2 < 15.89005 to the right, improve=0.5946574, (0 missing)
##
## Node number 53: 39 observations, complexity param=0.002118644
## predicted class=Y expected loss=0.4358974 P(node) =0.03823529
## class counts: 17 22
## probabilities: 0.436 0.564
## left son=106 (15 obs) right son=107 (24 obs)
## Primary splits:
## sprice.root2 < 18.70815 to the right, improve=0.4628205, (0 missing)
##
## Node number 54: 9 observations
## predicted class=N expected loss=0.3333333 P(node) =0.008823529
## class counts: 6 3
## probabilities: 0.667 0.333
##
## Node number 55: 49 observations
## predicted class=Y expected loss=0.3673469 P(node) =0.04803922
## class counts: 18 31
## probabilities: 0.367 0.633
##
## Node number 90: 95 observations, complexity param=0.002118644
## predicted class=N expected loss=0.2421053 P(node) =0.09313725
## class counts: 72 23
## probabilities: 0.758 0.242
## left son=180 (23 obs) right son=181 (72 obs)
## Primary splits:
## sprice.root2 < 10.23963 to the right, improve=1.460984, (0 missing)
##
## Node number 91: 33 observations, complexity param=0.002118644
## predicted class=N expected loss=0.4545455 P(node) =0.03235294
## class counts: 18 15
## probabilities: 0.545 0.455
## left son=182 (26 obs) right son=183 (7 obs)
## Primary splits:
## sprice.root2 < 11.81057 to the right, improve=1.198801, (0 missing)
##
## Node number 104: 7 observations
## predicted class=N expected loss=0.1428571 P(node) =0.006862745
## class counts: 6 1
## probabilities: 0.857 0.143
##
## Node number 105: 41 observations, complexity param=0.001059322
## predicted class=N expected loss=0.3658537 P(node) =0.04019608
## class counts: 26 15
## probabilities: 0.634 0.366
## left son=210 (32 obs) right son=211 (9 obs)
## Primary splits:
## sprice.root2 < 15.8106 to the left, improve=0.8299458, (0 missing)
##
## Node number 106: 15 observations
## predicted class=N expected loss=0.4666667 P(node) =0.01470588
## class counts: 8 7
## probabilities: 0.533 0.467
##
## Node number 107: 24 observations
## predicted class=Y expected loss=0.375 P(node) =0.02352941
## class counts: 9 15
## probabilities: 0.375 0.625
##
## Node number 180: 23 observations
## predicted class=N expected loss=0.08695652 P(node) =0.02254902
## class counts: 21 2
## probabilities: 0.913 0.087
##
## Node number 181: 72 observations, complexity param=0.002118644
## predicted class=N expected loss=0.2916667 P(node) =0.07058824
## class counts: 51 21
## probabilities: 0.708 0.292
## left son=362 (10 obs) right son=363 (62 obs)
## Primary splits:
## sprice.root2 < 5.417281 to the left, improve=0.8532258, (0 missing)
##
## Node number 182: 26 observations
## predicted class=N expected loss=0.3846154 P(node) =0.0254902
## class counts: 16 10
## probabilities: 0.615 0.385
##
## Node number 183: 7 observations
## predicted class=Y expected loss=0.2857143 P(node) =0.006862745
## class counts: 2 5
## probabilities: 0.286 0.714
##
## Node number 210: 32 observations
## predicted class=N expected loss=0.3125 P(node) =0.03137255
## class counts: 22 10
## probabilities: 0.688 0.312
##
## Node number 211: 9 observations
## predicted class=Y expected loss=0.4444444 P(node) =0.008823529
## class counts: 4 5
## probabilities: 0.444 0.556
##
## Node number 362: 10 observations
## predicted class=N expected loss=0.1 P(node) =0.009803922
## class counts: 9 1
## probabilities: 0.900 0.100
##
## Node number 363: 62 observations, complexity param=0.002118644
## predicted class=N expected loss=0.3225806 P(node) =0.06078431
## class counts: 42 20
## probabilities: 0.677 0.323
## left son=726 (55 obs) right son=727 (7 obs)
## Primary splits:
## sprice.root2 < 7.602942 to the right, improve=2.42145, (0 missing)
##
## Node number 726: 55 observations, complexity param=0.001059322
## predicted class=N expected loss=0.2727273 P(node) =0.05392157
## class counts: 40 15
## probabilities: 0.727 0.273
## left son=1452 (29 obs) right son=1453 (26 obs)
## Primary splits:
## sprice.root2 < 9.486569 to the left, improve=1.234627, (0 missing)
##
## Node number 727: 7 observations
## predicted class=Y expected loss=0.2857143 P(node) =0.006862745
## class counts: 2 5
## probabilities: 0.286 0.714
##
## Node number 1452: 29 observations
## predicted class=N expected loss=0.1724138 P(node) =0.02843137
## class counts: 24 5
## probabilities: 0.828 0.172
##
## Node number 1453: 26 observations, complexity param=0.001059322
## predicted class=N expected loss=0.3846154 P(node) =0.0254902
## class counts: 16 10
## probabilities: 0.615 0.385
## left son=2906 (19 obs) right son=2907 (7 obs)
## Primary splits:
## sprice.root2 < 9.565528 to the right, improve=0.6685946, (0 missing)
##
## Node number 2906: 19 observations
## predicted class=N expected loss=0.3157895 P(node) =0.01862745
## class counts: 13 6
## probabilities: 0.684 0.316
##
## Node number 2907: 7 observations
## predicted class=Y expected loss=0.4285714 P(node) =0.006862745
## class counts: 3 4
## probabilities: 0.429 0.571
##
## n= 1020
##
## node), split, n, loss, yval, (yprob)
## * denotes terminal node
##
## 1) root 1020 472 N (0.53725490 0.46274510)
## 2) biddable< 0.5 554 110 N (0.80144404 0.19855596)
## 4) sprice.root2>=19.90914 124 11 N (0.91129032 0.08870968) *
## 5) sprice.root2< 19.90914 430 99 N (0.76976744 0.23023256)
## 10) sprice.root2>=13.61047 263 51 N (0.80608365 0.19391635) *
## 11) sprice.root2< 13.61047 167 48 N (0.71257485 0.28742515)
## 22) sprice.root2< 13.41548 158 41 N (0.74050633 0.25949367)
## 44) sprice.root2>=12.54832 30 3 N (0.90000000 0.10000000) *
## 45) sprice.root2< 12.54832 128 38 N (0.70312500 0.29687500)
## 90) sprice.root2< 11.40153 95 23 N (0.75789474 0.24210526)
## 180) sprice.root2>=10.23963 23 2 N (0.91304348 0.08695652) *
## 181) sprice.root2< 10.23963 72 21 N (0.70833333 0.29166667)
## 362) sprice.root2< 5.417281 10 1 N (0.90000000 0.10000000) *
## 363) sprice.root2>=5.417281 62 20 N (0.67741935 0.32258065)
## 726) sprice.root2>=7.602942 55 15 N (0.72727273 0.27272727)
## 1452) sprice.root2< 9.486569 29 5 N (0.82758621 0.17241379) *
## 1453) sprice.root2>=9.486569 26 10 N (0.61538462 0.38461538)
## 2906) sprice.root2>=9.565528 19 6 N (0.68421053 0.31578947) *
## 2907) sprice.root2< 9.565528 7 3 Y (0.42857143 0.57142857) *
## 727) sprice.root2< 7.602942 7 2 Y (0.28571429 0.71428571) *
## 91) sprice.root2>=11.40153 33 15 N (0.54545455 0.45454545)
## 182) sprice.root2>=11.81057 26 10 N (0.61538462 0.38461538) *
## 183) sprice.root2< 11.81057 7 2 Y (0.28571429 0.71428571) *
## 23) sprice.root2>=13.41548 9 2 Y (0.22222222 0.77777778) *
## 3) biddable>=0.5 466 104 Y (0.22317597 0.77682403)
## 6) sprice.root2>=11.83195 164 76 N (0.53658537 0.46341463)
## 12) sprice.root2>=20.33187 19 4 N (0.78947368 0.21052632) *
## 13) sprice.root2< 20.33187 145 72 N (0.50344828 0.49655172)
## 26) sprice.root2>=13.89077 87 38 N (0.56321839 0.43678161)
## 52) sprice.root2< 16.50717 48 16 N (0.66666667 0.33333333)
## 104) sprice.root2>=15.89005 7 1 N (0.85714286 0.14285714) *
## 105) sprice.root2< 15.89005 41 15 N (0.63414634 0.36585366)
## 210) sprice.root2< 15.8106 32 10 N (0.68750000 0.31250000) *
## 211) sprice.root2>=15.8106 9 4 Y (0.44444444 0.55555556) *
## 53) sprice.root2>=16.50717 39 17 Y (0.43589744 0.56410256)
## 106) sprice.root2>=18.70815 15 7 N (0.53333333 0.46666667) *
## 107) sprice.root2< 18.70815 24 9 Y (0.37500000 0.62500000) *
## 27) sprice.root2< 13.89077 58 24 Y (0.41379310 0.58620690)
## 54) sprice.root2< 12.24684 9 3 N (0.66666667 0.33333333) *
## 55) sprice.root2>=12.24684 49 18 Y (0.36734694 0.63265306) *
## 7) sprice.root2< 11.83195 302 16 Y (0.05298013 0.94701987) *
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.1X1.cp.0.rpart.N
## 1 N 484
## 2 Y 107
## sold.fctr.predict.Max.cor.Y.rcv.1X1.cp.0.rpart.Y
## 1 64
## 2 365
## Prediction
## Reference N Y
## N 484 64
## Y 107 365
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 8.323529e-01 6.606905e-01 8.079872e-01 8.547824e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 4.198532e-88 1.318969e-03
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.1X1.cp.0.rpart.N
## 1 N 327
## 2 Y 93
## sold.fctr.predict.Max.cor.Y.rcv.1X1.cp.0.rpart.Y
## 1 120
## 2 290
## Prediction
## Reference N Y
## N 327 120
## Y 93 290
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.433735e-01 4.862697e-01 7.122254e-01 7.727821e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 4.300069e-34 7.483233e-02
## id feats max.nTuningRuns
## 1 Max.cor.Y.rcv.1X1.cp.0.rpart biddable,sprice.root2 0
## min.elapsedtime.everything min.elapsedtime.final max.AUCpROC.fit
## 1 0.75 0.012 0.8281424
## max.Sens.fit max.Spec.fit max.AUCROCR.fit opt.prob.threshold.fit
## 1 0.8978102 0.7584746 0.8816304 0.4
## max.f.score.fit max.Accuracy.fit max.AccuracyLower.fit
## 1 0.8102109 0.8323529 0.8079872
## max.AccuracyUpper.fit max.Kappa.fit max.AUCpROC.OOB max.Sens.OOB
## 1 0.8547824 0.6606905 0.750495 0.8456376
## max.Spec.OOB max.AUCROCR.OOB opt.prob.threshold.OOB max.f.score.OOB
## 1 0.6553525 0.7934621 0.3 0.7313997
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.7433735 0.7122254 0.7727821
## max.Kappa.OOB
## 1 0.4862697#stop(here"); glb_to_sav(); all.equal(glb_models_df, sav_models_df)
# if (glb_is_regression || glb_is_binomial) # For multinomials this model will be run next by default
ret_lst <- myfit_mdl(mdl_specs_lst=myinit_mdl_specs_lst(mdl_specs_lst=list(
id.prefix="Max.cor.Y",
type=glb_model_type, trainControl.method="repeatedcv",
trainControl.number=glb_rcv_n_folds, trainControl.repeats=glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method="rpart")),
indep_vars=max_cor_y_x_vars, rsp_var=glb_rsp_var,
fit_df=glb_fitobs_df, OOB_df=glb_OOBobs_df)## [1] "fitting model: Max.cor.Y.rpart"
## [1] " indep_vars: biddable,sprice.root2"
## + Fold1.Rep1: cp=0.003531
## - Fold1.Rep1: cp=0.003531
## + Fold2.Rep1: cp=0.003531
## - Fold2.Rep1: cp=0.003531
## + Fold3.Rep1: cp=0.003531
## - Fold3.Rep1: cp=0.003531
## + Fold1.Rep2: cp=0.003531
## - Fold1.Rep2: cp=0.003531
## + Fold2.Rep2: cp=0.003531
## - Fold2.Rep2: cp=0.003531
## + Fold3.Rep2: cp=0.003531
## - Fold3.Rep2: cp=0.003531
## + Fold1.Rep3: cp=0.003531
## - Fold1.Rep3: cp=0.003531
## + Fold2.Rep3: cp=0.003531
## - Fold2.Rep3: cp=0.003531
## + Fold3.Rep3: cp=0.003531
## - Fold3.Rep3: cp=0.003531
## Aggregating results
## Selecting tuning parameters
## Fitting cp = 0.00636 on full training set
## Call:
## rpart(formula = .outcome ~ ., control = list(minsplit = 20, minbucket = 7,
## cp = 0, maxcompete = 4, maxsurrogate = 5, usesurrogate = 2,
## surrogatestyle = 0, maxdepth = 30, xval = 0))
## n= 1020
##
## CP nsplit rel error
## 1 0.546610169 0 1.0000000
## 2 0.025423729 1 0.4533898
## 3 0.010593220 2 0.4279661
## 4 0.006355932 5 0.3961864
##
## Variable importance
## biddable sprice.root2
## 56 44
##
## Node number 1: 1020 observations, complexity param=0.5466102
## predicted class=N expected loss=0.4627451 P(node) =1
## class counts: 548 472
## probabilities: 0.537 0.463
## left son=2 (554 obs) right son=3 (466 obs)
## Primary splits:
## biddable < 0.5 to the left, improve=169.2715, (0 missing)
## sprice.root2 < 10.23963 to the right, improve=133.6444, (0 missing)
## Surrogate splits:
## sprice.root2 < 12.26518 to the right, agree=0.751, adj=0.455, (0 split)
##
## Node number 2: 554 observations
## predicted class=N expected loss=0.198556 P(node) =0.5431373
## class counts: 444 110
## probabilities: 0.801 0.199
##
## Node number 3: 466 observations, complexity param=0.02542373
## predicted class=Y expected loss=0.223176 P(node) =0.4568627
## class counts: 104 362
## probabilities: 0.223 0.777
## left son=6 (164 obs) right son=7 (302 obs)
## Primary splits:
## sprice.root2 < 11.83195 to the right, improve=49.71379, (0 missing)
##
## Node number 6: 164 observations, complexity param=0.01059322
## predicted class=N expected loss=0.4634146 P(node) =0.1607843
## class counts: 88 76
## probabilities: 0.537 0.463
## left son=12 (19 obs) right son=13 (145 obs)
## Primary splits:
## sprice.root2 < 20.33187 to the right, improve=2.748634, (0 missing)
##
## Node number 7: 302 observations
## predicted class=Y expected loss=0.05298013 P(node) =0.2960784
## class counts: 16 286
## probabilities: 0.053 0.947
##
## Node number 12: 19 observations
## predicted class=N expected loss=0.2105263 P(node) =0.01862745
## class counts: 15 4
## probabilities: 0.789 0.211
##
## Node number 13: 145 observations, complexity param=0.01059322
## predicted class=N expected loss=0.4965517 P(node) =0.1421569
## class counts: 73 72
## probabilities: 0.503 0.497
## left son=26 (87 obs) right son=27 (58 obs)
## Primary splits:
## sprice.root2 < 13.89077 to the right, improve=1.554023, (0 missing)
##
## Node number 26: 87 observations, complexity param=0.01059322
## predicted class=N expected loss=0.4367816 P(node) =0.08529412
## class counts: 49 38
## probabilities: 0.563 0.437
## left son=52 (48 obs) right son=53 (39 obs)
## Primary splits:
## sprice.root2 < 16.50717 to the left, improve=2.291777, (0 missing)
##
## Node number 27: 58 observations
## predicted class=Y expected loss=0.4137931 P(node) =0.05686275
## class counts: 24 34
## probabilities: 0.414 0.586
##
## Node number 52: 48 observations
## predicted class=N expected loss=0.3333333 P(node) =0.04705882
## class counts: 32 16
## probabilities: 0.667 0.333
##
## Node number 53: 39 observations
## predicted class=Y expected loss=0.4358974 P(node) =0.03823529
## class counts: 17 22
## probabilities: 0.436 0.564
##
## n= 1020
##
## node), split, n, loss, yval, (yprob)
## * denotes terminal node
##
## 1) root 1020 472 N (0.53725490 0.46274510)
## 2) biddable< 0.5 554 110 N (0.80144404 0.19855596) *
## 3) biddable>=0.5 466 104 Y (0.22317597 0.77682403)
## 6) sprice.root2>=11.83195 164 76 N (0.53658537 0.46341463)
## 12) sprice.root2>=20.33187 19 4 N (0.78947368 0.21052632) *
## 13) sprice.root2< 20.33187 145 72 N (0.50344828 0.49655172)
## 26) sprice.root2>=13.89077 87 38 N (0.56321839 0.43678161)
## 52) sprice.root2< 16.50717 48 16 N (0.66666667 0.33333333) *
## 53) sprice.root2>=16.50717 39 17 Y (0.43589744 0.56410256) *
## 27) sprice.root2< 13.89077 58 24 Y (0.41379310 0.58620690) *
## 7) sprice.root2< 11.83195 302 16 Y (0.05298013 0.94701987) *
## sold.fctr sold.fctr.predict.Max.cor.Y.rpart.N
## 1 N 491
## 2 Y 130
## sold.fctr.predict.Max.cor.Y.rpart.Y
## 1 57
## 2 342
## Prediction
## Reference N Y
## N 491 57
## Y 130 342
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 8.166667e-01 6.272891e-01 7.915283e-01 8.399619e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 2.183067e-78 1.400662e-07
## sold.fctr sold.fctr.predict.Max.cor.Y.rpart.N
## 1 N 368
## 2 Y 112
## sold.fctr.predict.Max.cor.Y.rpart.Y
## 1 79
## 2 271
## Prediction
## Reference N Y
## N 368 79
## Y 112 271
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.698795e-01 5.341327e-01 7.397113e-01 7.981170e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 1.766677e-43 2.058893e-02
## id feats max.nTuningRuns
## 1 Max.cor.Y.rpart biddable,sprice.root2 5
## min.elapsedtime.everything min.elapsedtime.final max.AUCpROC.fit
## 1 1.632 0.013 0.8102808
## max.Sens.fit max.Spec.fit max.AUCROCR.fit opt.prob.threshold.fit
## 1 0.8959854 0.7245763 0.8387008 0.5
## max.f.score.fit max.Accuracy.fit max.AccuracyLower.fit
## 1 0.7853042 0.7964073 0.7915283
## max.AccuracyUpper.fit max.Kappa.fit max.AUCpROC.OOB max.Sens.OOB
## 1 0.8399619 0.5852116 0.7445254 0.8545861
## max.Spec.OOB max.AUCROCR.OOB opt.prob.threshold.OOB max.f.score.OOB
## 1 0.6344648 0.7855854 0.3 0.739427
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.7698795 0.7397113 0.798117
## max.Kappa.OOB max.AccuracySD.fit max.KappaSD.fit
## 1 0.5341327 0.02006899 0.04005915if (!is.null(glb_date_vars) &&
(sum(grepl(paste(glb_date_vars, "\\.day\\.minutes\\.poly\\.", sep=""),
names(glb_allobs_df))) > 0)) {
# ret_lst <- myfit_mdl(mdl_id="Max.cor.Y.TmSrs.poly1",
# model_method=ifelse(glb_is_regression, "lm",
# ifelse(glb_is_binomial, "glm", "rpart")),
# model_type=glb_model_type,
# indep_vars_vctr=c(max_cor_y_x_vars, paste0(glb_date_vars, ".day.minutes")),
# rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
# fit_df=glb_fitobs_df, OOB_df=glb_OOBobs_df,
# n_cv_folds=glb_rcv_n_folds, tune_models_df=NULL)
#
ret_lst <- myfit_mdl(mdl_id="Max.cor.Y.TmSrs.poly",
model_method=ifelse(glb_is_regression, "lm",
ifelse(glb_is_binomial, "glm", "rpart")),
model_type=glb_model_type,
indep_vars_vctr=c(max_cor_y_x_vars,
grep(paste(glb_date_vars, "\\.day\\.minutes\\.poly\\.", sep=""),
names(glb_allobs_df), value=TRUE)),
rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
fit_df=glb_fitobs_df, OOB_df=glb_OOBobs_df,
n_cv_folds=glb_rcv_n_folds, tune_models_df=NULL)
}
# Interactions.High.cor.Y
if (length(int_feats <- setdiff(setdiff(unique(glb_feats_df$cor.high.X), NA),
subset(glb_feats_df, nzv)$id)) > 0) {
ret_lst <- myfit_mdl(mdl_specs_lst=myinit_mdl_specs_lst(mdl_specs_lst=list(
id.prefix="Interact.High.cor.Y",
type=glb_model_type, trainControl.method="repeatedcv",
trainControl.number=glb_rcv_n_folds, trainControl.repeats=glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method="glmnet")),
indep_vars=c(max_cor_y_x_vars, paste(max_cor_y_x_vars[1], int_feats, sep=":")),
rsp_var=glb_rsp_var,
fit_df=glb_fitobs_df, OOB_df=glb_OOBobs_df)
} ## [1] "fitting model: Interact.High.cor.Y.glmnet"
## [1] " indep_vars: biddable,sprice.root2,biddable:sprice.log10,biddable:D.terms.post.stop.n.log,biddable:startprice.dcm2.is9,biddable:D.weight.post.stem.sum,biddable:D.chrs.n.log,biddable:cellular.fctr,biddable:D.terms.post.stem.n.log,biddable:sprice.root2"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0.325, lambda = 0.00124 on full training set
## Length Class Mode
## a0 100 -none- numeric
## beta 1100 dgCMatrix S4
## df 100 -none- numeric
## dim 2 -none- numeric
## lambda 100 -none- numeric
## dev.ratio 100 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 11 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept) biddable
## 0.01092205 4.74489316
## sprice.root2 biddable:D.chrs.n.log
## -0.08925084 0.07353070
## biddable:D.terms.post.stop.n.log biddable:D.weight.post.stem.sum
## -0.03122283 -0.14511418
## biddable:cellular.fctr1 biddable:cellular.fctrUnknown
## -0.13728156 -1.05190220
## biddable:sprice.log10 biddable:sprice.root2
## -0.10196500 -0.15254153
## biddable:startprice.dcm2.is9
## -0.04646263
## [1] "max lambda < lambdaOpt:"
## (Intercept) biddable
## -0.002342312 4.814280953
## sprice.root2 biddable:D.chrs.n.log
## -0.088443437 0.099967006
## biddable:D.terms.post.stop.n.log biddable:D.weight.post.stem.sum
## -0.057642263 -0.156438317
## biddable:cellular.fctr1 biddable:cellular.fctrUnknown
## -0.141662526 -1.060718650
## biddable:sprice.log10 biddable:sprice.root2
## -0.113943290 -0.152574470
## biddable:startprice.dcm2.is9
## -0.056816165
## sold.fctr sold.fctr.predict.Interact.High.cor.Y.glmnet.N
## 1 N 467
## 2 Y 124
## sold.fctr.predict.Interact.High.cor.Y.glmnet.Y
## 1 81
## 2 348
## Prediction
## Reference N Y
## N 467 81
## Y 124 348
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.990196e-01 5.932255e-01 7.730893e-01 8.232111e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 2.536091e-68 3.352638e-03
## sold.fctr sold.fctr.predict.Interact.High.cor.Y.glmnet.N
## 1 N 349
## 2 Y 93
## sold.fctr.predict.Interact.High.cor.Y.glmnet.Y
## 1 98
## 2 290
## Prediction
## Reference N Y
## N 349 98
## Y 93 290
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.698795e-01 5.374385e-01 7.397113e-01 7.981170e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 1.766677e-43 7.722525e-01
## id
## 1 Interact.High.cor.Y.glmnet
## feats
## 1 biddable,sprice.root2,biddable:sprice.log10,biddable:D.terms.post.stop.n.log,biddable:startprice.dcm2.is9,biddable:D.weight.post.stem.sum,biddable:D.chrs.n.log,biddable:cellular.fctr,biddable:D.terms.post.stem.n.log,biddable:sprice.root2
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 25 2.747 0.1
## max.AUCpROC.fit max.Sens.fit max.Spec.fit max.AUCROCR.fit
## 1 0.7970973 0.8886861 0.7055085 0.863937
## opt.prob.threshold.fit max.f.score.fit max.Accuracy.fit
## 1 0.4 0.772475 0.8026098
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.7730893 0.8232111 0.5988972
## max.AUCpROC.OOB max.Sens.OOB max.Spec.OOB max.AUCROCR.OOB
## 1 0.7551533 0.8680089 0.6422977 0.8189
## opt.prob.threshold.OOB max.f.score.OOB max.Accuracy.OOB
## 1 0.3 0.7522698 0.7698795
## max.AccuracyLower.OOB max.AccuracyUpper.OOB max.Kappa.OOB
## 1 0.7397113 0.798117 0.5374385
## max.AccuracySD.fit max.KappaSD.fit
## 1 0.01796977 0.03705653# Low.cor.X
# if (glb_is_classification && glb_is_binomial)
# indep_vars_vctr <- subset(glb_feats_df, is.na(cor.high.X) &
# is.ConditionalX.y &
# (exclude.as.feat != 1))[, "id"] else
indep_vars <- subset(glb_feats_df, is.na(cor.high.X) & !nzv &
(exclude.as.feat != 1))[, "id"]
indep_vars <- myadjust_interaction_feats(indep_vars)
ret_lst <- myfit_mdl(mdl_specs_lst=myinit_mdl_specs_lst(mdl_specs_lst=list(
id.prefix="Low.cor.X",
type=glb_model_type, trainControl.method="repeatedcv",
trainControl.number=glb_rcv_n_folds, trainControl.repeats=glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method="glmnet")),
indep_vars=indep_vars, rsp_var=glb_rsp_var,
fit_df=glb_fitobs_df, OOB_df=glb_OOBobs_df)## [1] "fitting model: Low.cor.X.glmnet"
## [1] " indep_vars: biddable,spdiff.cut.fctr,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,storage.fctr,D.chrs.pnct11.n.log,startprice.dgt2.is9,color.fctr,prdl.my.fctr,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.root2,prdl.my.fctr:.clusterid.fctr"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0.775, lambda = 0.00124 on full training set
## Length Class Mode
## a0 86 -none- numeric
## beta 5332 dgCMatrix S4
## df 86 -none- numeric
## dim 2 -none- numeric
## lambda 86 -none- numeric
## dev.ratio 86 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 62 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept)
## 0.28483910
## .rnorm
## -0.02167279
## D.chrs.pnct11.n.log
## 0.40659853
## D.ratio.weight.sum.wrds.n
## -0.11673547
## D.terms.post.stop.n.log
## -0.04986918
## D.weight.sum.stem.stop.Ratio
## -1.64844144
## D.wrds.stop.n.log
## -0.06590812
## biddable
## 2.44322986
## cellular.fctrUnknown
## -1.42750817
## color.fctrGold
## 0.05191392
## color.fctrSpace Gray
## -0.31389140
## color.fctrUnknown
## 0.32139639
## color.fctrWhite
## -0.14084948
## condition.fctrFor parts or not working
## -0.63820054
## condition.fctrManufacturer refurbished
## -1.03703740
## condition.fctrNew other (see details)
## 0.53706086
## condition.fctrSeller refurbished
## -0.92779447
## prdl.my.fctriPad1
## -0.73094246
## prdl.my.fctriPad2
## -0.50827810
## prdl.my.fctriPad3
## 0.09236029
## prdl.my.fctriPad4
## 1.02122755
## prdl.my.fctriPadAir
## 1.27964442
## prdl.my.fctriPadAir2
## 1.73963473
## prdl.my.fctriPadmini
## -0.45817873
## prdl.my.fctriPadmini2
## 1.07559291
## prdl.my.fctriPadmini3
## 0.58081997
## spdiff.cut.fctr(-100,-10]
## 2.03718267
## spdiff.cut.fctr(-10,-1]
## 3.78016051
## spdiff.cut.fctr(-1,0]
## 4.91900841
## spdiff.cut.fctr(0,1]
## 4.02364901
## spdiff.cut.fctr(1,10]
## 4.05473201
## spdiff.cut.fctr(10,100]
## 3.51590467
## spdiff.cut.fctr(100,1e+03]
## 0.83208831
## sprice.root2
## -0.17520239
## startprice.dcm2.is9
## -0.37526422
## startprice.dgt1.is9
## -0.12236933
## startprice.dgt2.is9
## 0.12473613
## storage.fctr16
## -0.10386074
## storage.fctr64
## 0.19776093
## storage.fctrUnknown
## 1.37356185
## prdl.my.fctrUnknown:.clusterid.fctr2
## 0.75846077
## prdl.my.fctriPad1:.clusterid.fctr2
## -0.47363807
## prdl.my.fctriPad3:.clusterid.fctr2
## 0.23980635
## prdl.my.fctriPad4:.clusterid.fctr2
## -1.15493595
## prdl.my.fctriPadAir:.clusterid.fctr2
## -0.30219482
## prdl.my.fctriPadmini2:.clusterid.fctr2
## -0.59647290
## prdl.my.fctrUnknown:.clusterid.fctr3
## -0.71399278
## prdl.my.fctriPad1:.clusterid.fctr3
## -0.74280799
## prdl.my.fctriPad4:.clusterid.fctr3
## -3.99203955
## [1] "max lambda < lambdaOpt:"
## (Intercept)
## 0.30739060
## .rnorm
## -0.02307603
## D.chrs.pnct11.n.log
## 0.41194798
## D.ratio.weight.sum.wrds.n
## -0.11841828
## D.terms.post.stop.n.log
## -0.04857498
## D.weight.sum.stem.stop.Ratio
## -1.70659855
## D.wrds.stop.n.log
## -0.07092480
## biddable
## 2.45597112
## cellular.fctrUnknown
## -1.44850328
## color.fctrGold
## 0.06040873
## color.fctrSpace Gray
## -0.31733631
## color.fctrUnknown
## 0.32580290
## color.fctrWhite
## -0.14400741
## condition.fctrFor parts or not working
## -0.64810440
## condition.fctrManufacturer refurbished
## -1.05310177
## condition.fctrNew other (see details)
## 0.54221686
## condition.fctrSeller refurbished
## -0.94083418
## prdl.my.fctriPad1
## -0.72962954
## prdl.my.fctriPad2
## -0.50283877
## prdl.my.fctriPad3
## 0.10518937
## prdl.my.fctriPad4
## 1.05514873
## prdl.my.fctriPadAir
## 1.31094762
## prdl.my.fctriPadAir2
## 1.76759552
## prdl.my.fctriPadmini
## -0.45080046
## prdl.my.fctriPadmini2
## 1.10294883
## prdl.my.fctriPadmini3
## 0.60534029
## spdiff.cut.fctr(-100,-10]
## 2.07034040
## spdiff.cut.fctr(-10,-1]
## 3.81948497
## spdiff.cut.fctr(-1,0]
## 4.97713395
## spdiff.cut.fctr(0,1]
## 4.07618809
## spdiff.cut.fctr(1,10]
## 4.09721806
## spdiff.cut.fctr(10,100]
## 3.55524679
## spdiff.cut.fctr(100,1e+03]
## 0.86203115
## sprice.root2
## -0.17628994
## startprice.dcm2.is9
## -0.38089328
## startprice.dgt1.is9
## -0.12199526
## startprice.dgt2.is9
## 0.12986601
## storage.fctr16
## -0.10644192
## storage.fctr64
## 0.19978492
## storage.fctrUnknown
## 1.40216390
## prdl.my.fctrUnknown:.clusterid.fctr2
## 0.77627886
## prdl.my.fctriPad1:.clusterid.fctr2
## -0.48242024
## prdl.my.fctriPad3:.clusterid.fctr2
## 0.24204519
## prdl.my.fctriPad4:.clusterid.fctr2
## -1.18680995
## prdl.my.fctriPadAir:.clusterid.fctr2
## -0.31435860
## prdl.my.fctriPadmini2:.clusterid.fctr2
## -0.60900002
## prdl.my.fctrUnknown:.clusterid.fctr3
## -0.72474486
## prdl.my.fctriPad1:.clusterid.fctr3
## -0.75028690
## prdl.my.fctriPad4:.clusterid.fctr3
## -4.10529031
## sold.fctr sold.fctr.predict.Low.cor.X.glmnet.N
## 1 N 493
## 2 Y 85
## sold.fctr.predict.Low.cor.X.glmnet.Y
## 1 55
## 2 387
## Prediction
## Reference N Y
## N 493 55
## Y 85 387
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 8.627451e-01 7.227357e-01 8.400893e-01 8.832827e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 3.342931e-109 1.424808e-02
## sold.fctr sold.fctr.predict.Low.cor.X.glmnet.N
## 1 N 313
## 2 Y 58
## sold.fctr.predict.Low.cor.X.glmnet.Y
## 1 134
## 2 325
## Prediction
## Reference N Y
## N 313 134
## Y 58 325
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.686747e-01 5.411011e-01 7.384586e-01 7.969687e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 5.056117e-43 6.209575e-08
## id
## 1 Low.cor.X.glmnet
## feats
## 1 biddable,spdiff.cut.fctr,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,storage.fctr,D.chrs.pnct11.n.log,startprice.dgt2.is9,color.fctr,prdl.my.fctr,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.root2,prdl.my.fctr:.clusterid.fctr
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 25 3.129 0.121
## max.AUCpROC.fit max.Sens.fit max.Spec.fit max.AUCROCR.fit
## 1 0.8597751 0.899635 0.8199153 0.9329264
## opt.prob.threshold.fit max.f.score.fit max.Accuracy.fit
## 1 0.5 0.8468271 0.8349657
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.8400893 0.8832827 0.6666179
## max.AUCpROC.OOB max.Sens.OOB max.Spec.OOB max.AUCROCR.OOB
## 1 0.7801532 0.8187919 0.7415144 0.8727052
## opt.prob.threshold.OOB max.f.score.OOB max.Accuracy.OOB
## 1 0.3 0.7719715 0.7686747
## max.AccuracyLower.OOB max.AccuracyUpper.OOB max.Kappa.OOB
## 1 0.7384586 0.7969687 0.5411011
## max.AccuracySD.fit max.KappaSD.fit
## 1 0.009378705 0.01938402rm(ret_lst)
glb_chunks_df <- myadd_chunk(glb_chunks_df, "fit.models", major.inc=FALSE)## label step_major step_minor label_minor bgn end elapsed
## 10 fit.models 6 0 0 342.849 409.618 66.769
## 11 fit.models 6 1 1 409.619 NA NAfit.models_1_chunk_df <- myadd_chunk(NULL, "fit.models_1_bgn", label.minor="setup")## label step_major step_minor label_minor bgn end elapsed
## 1 fit.models_1_bgn 1 0 setup 423.992 NA NA#stop(here"); glb_to_sav(); all.equal(glb_models_df, sav_models_df)
topindep_var <- NULL; interact_vars <- NULL;
for (mdl_id_pfx in names(glb_mdl_family_lst)) {
fit.models_1_chunk_df <- myadd_chunk(fit.models_1_chunk_df,
paste0("fit.models_1_", mdl_id_pfx), major.inc = TRUE, label.minor = "setup")
indep_vars <- NULL;
if (grepl("\\.Interact", mdl_id_pfx)) {
if (is.null(topindep_var) && is.null(interact_vars)) {
# select best glmnet model upto now
dsp_models_df <- orderBy(model_sel_frmla <- get_model_sel_frmla(),
glb_models_df)
dsp_models_df <- subset(dsp_models_df, grepl(".glmnet", id, fixed=TRUE))
bst_mdl_id <- dsp_models_df$id[1]
mdl_id_pfx <-
paste(c(head(unlist(strsplit(bst_mdl_id, "[.]")), -1), "Interact"),
collapse=".")
# select most importance feature
if (is.null(bst_featsimp_df <-
myget_feats_importance(glb_models_lst[[bst_mdl_id]]))) {
warning("Base model for RFE.Interact: ", bst_mdl_id,
" has no important features")
next
}
topindep_ix <- 1
while (is.null(topindep_var) && (topindep_ix <= nrow(bst_featsimp_df))) {
topindep_var <- row.names(bst_featsimp_df)[topindep_ix]
if (grepl(".fctr", topindep_var, fixed=TRUE))
topindep_var <-
paste0(unlist(strsplit(topindep_var, ".fctr"))[1], ".fctr")
if (topindep_var %in% names(glb_interaction_only_feats_lst)) {
topindep_var <- NULL; topindep_ix <- topindep_ix + 1
} else break
}
# select features with importance > max(10, importance of .rnorm) & is not highest
# combine factor dummy features to just the factor feature
if (length(pos_rnorm <-
grep(".rnorm", row.names(bst_featsimp_df), fixed=TRUE)) > 0)
imp_rnorm <- bst_featsimp_df[pos_rnorm, 1] else
imp_rnorm <- NA
importance_cutoff <- max(10, imp_rnorm, na.rm=TRUE)
interact_vars <-
tail(row.names(subset(bst_featsimp_df,
importance > importance_cutoff)), -1)
if (length(interact_vars) > 0) {
interact_vars <-
myadjust_interaction_feats(myextract_actual_feats(interact_vars))
interact_vars <-
interact_vars[!grepl(topindep_var, interact_vars, fixed=TRUE)]
}
### bid0_sp only
# interact_vars <- c(
# "biddable", "D.ratio.sum.TfIdf.wrds.n", "D.TfIdf.sum.stem.stop.Ratio", "D.sum.TfIdf",
# "D.TfIdf.sum.post.stop", "D.TfIdf.sum.post.stem", "D.ratio.wrds.stop.n.wrds.n", "D.chrs.uppr.n.log",
# "D.chrs.n.log", "color.fctr"
# # , "condition.fctr", "prdl.my.descr.fctr"
# )
# interact_vars <- setdiff(interact_vars, c("startprice.dgt2.is9", "color.fctr"))
###
indep_vars <- myextract_actual_feats(row.names(bst_featsimp_df))
indep_vars <- setdiff(indep_vars, topindep_var)
if (length(interact_vars) > 0) {
indep_vars <-
setdiff(indep_vars, myextract_actual_feats(interact_vars))
indep_vars <- c(indep_vars,
paste(topindep_var, setdiff(interact_vars, topindep_var),
sep = "*"))
} else indep_vars <- union(indep_vars, topindep_var)
}
}
if (is.null(indep_vars))
indep_vars <- glb_mdl_feats_lst[[mdl_id_pfx]]
if (is.null(indep_vars) && grepl("RFE\\.", mdl_id_pfx))
indep_vars <- myextract_actual_feats(predictors(rfe_fit_results))
if (is.null(indep_vars))
indep_vars <- subset(glb_feats_df, !nzv & (exclude.as.feat != 1))[, "id"]
indep_vars <- myadjust_interaction_feats(indep_vars)
if (grepl("\\.Interact", mdl_id_pfx)) {
# if (method != tail(unlist(strsplit(bst_mdl_id, "[.]")), 1)) next
if (is.null(glb_mdl_family_lst[[mdl_id_pfx]])) {
if (!is.null(glb_mdl_family_lst[["Best.Interact"]]))
glb_mdl_family_lst[[mdl_id_pfx]] <-
glb_mdl_family_lst[["Best.Interact"]]
}
}
if (!is.null(glbObsFitOutliers[[mdl_id_pfx]])) {
fitobs_df <- glb_fitobs_df[!(glb_fitobs_df[, glb_id_var] %in%
glbObsFitOutliers[[mdl_id_pfx]]), ]
} else fitobs_df <- glb_fitobs_df
if (is.null(glb_mdl_family_lst[[mdl_id_pfx]]))
mdl_methods <- glbMdlMethods else
mdl_methods <- glb_mdl_family_lst[[mdl_id_pfx]]
for (method in mdl_methods) {
if (method %in% c("rpart", "rf")) {
# rpart: fubar's the tree
# rf: skip the scenario w/ .rnorm for speed
indep_vars <- setdiff(indep_vars, c(".rnorm"))
#mdl_id <- paste0(mdl_id_pfx, ".no.rnorm")
}
fit.models_1_chunk_df <- myadd_chunk(fit.models_1_chunk_df,
paste0("fit.models_1_", mdl_id_pfx), major.inc = FALSE,
label.minor = method)
ret_lst <-
myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = mdl_id_pfx,
type = glb_model_type, tune.df = glb_tune_models_df,
trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds,
trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
#trainControl.allowParallel = FALSE,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = method)),
indep_vars = indep_vars, rsp_var = glb_rsp_var,
fit_df = fitobs_df, OOB_df = glb_OOBobs_df)
}
}## label step_major step_minor label_minor bgn end
## 1 fit.models_1_bgn 1 0 setup 423.992 424.003
## 2 fit.models_1_RFE.X 2 0 setup 424.003 NA
## elapsed
## 1 0.011
## 2 NA
## label step_major step_minor label_minor bgn end
## 2 fit.models_1_RFE.X 2 0 setup 424.003 424.011
## 3 fit.models_1_RFE.X 2 1 glmnet 424.011 NA
## elapsed
## 2 0.008
## 3 NA
## [1] "fitting model: RFE.X.glmnet"
## [1] " indep_vars: sprice.d20nexp,sprice.log10,sprice.root2,biddable,spdiff.cut.fctr,cellular.fctr,prdl.my.fctr,D.ratio.wrds.stop.n.wrds.n,storage.fctr,condition.fctr,D.weight.sum.stem.stop.Ratio,color.fctr,startprice.dcm2.is9,.rnorm,D.chrs.pnct11.n.log,D.wrds.stop.n.log,startprice.dcm1.is9,startprice.dgt2.is9,D.ratio.weight.sum.wrds.n,D.chrs.pnct13.n.log,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,D.wrds.n.log,D.chrs.uppr.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,D.terms.post.stem.n.log,D.chrs.n.log,startprice.dgt1.is9,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0.1, lambda = 0.000267 on full training set## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst
## = list(id.prefix = mdl_id_pfx, : model's bestTune found at an extreme of
## tuneGrid for parameter: alpha## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst
## = list(id.prefix = mdl_id_pfx, : model's bestTune found at an extreme of
## tuneGrid for parameter: lambda
## Length Class Mode
## a0 100 -none- numeric
## beta 9300 dgCMatrix S4
## df 100 -none- numeric
## dim 2 -none- numeric
## lambda 100 -none- numeric
## dev.ratio 100 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 93 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept)
## 5.9298599950
## .rnorm
## -0.0367487387
## D.chrs.n.log
## -0.2852958123
## D.chrs.pnct11.n.log
## 0.4714450805
## D.chrs.pnct13.n.log
## -0.1653784579
## D.chrs.uppr.n.log
## -0.0116026565
## D.ratio.weight.sum.wrds.n
## -0.0593199934
## D.ratio.wrds.stop.n.wrds.n
## -2.9071388332
## D.terms.post.stem.n.log
## -0.2864766681
## D.terms.post.stop.n.log
## -0.9265313808
## D.weight.post.stem.sum
## -0.1209650771
## D.weight.sum
## -0.1180951219
## D.weight.sum.stem.stop.Ratio
## -2.9013605803
## D.wrds.n.log
## 1.2607015439
## D.wrds.stop.n.log
## 0.1152103538
## D.wrds.unq.n.log
## -0.1770678274
## biddable
## 2.5343114368
## cellular.fctrUnknown
## -0.7922522457
## color.fctrGold
## 0.0871372337
## color.fctrSpace Gray
## -0.3474018561
## color.fctrUnknown
## 0.3560862355
## color.fctrWhite
## -0.2253657549
## condition.fctrFor parts or not working
## -0.7365978083
## condition.fctrManufacturer refurbished
## -1.1969859100
## condition.fctrNew
## 0.0297417566
## condition.fctrNew other (see details)
## 0.7229925582
## condition.fctrSeller refurbished
## -1.0271638293
## prdl.my.fctriPad1
## -0.6856562063
## prdl.my.fctriPad2
## -0.4622200902
## prdl.my.fctriPad3
## 0.1837795731
## prdl.my.fctriPad4
## 1.4053112779
## prdl.my.fctriPadAir
## 1.5486082406
## prdl.my.fctriPadAir2
## 1.9064724435
## prdl.my.fctriPadmini
## -0.4755225796
## prdl.my.fctriPadmini2
## 1.3324979842
## prdl.my.fctriPadmini3
## 0.7573021470
## spdiff.cut.fctr(-100,-10]
## 2.3577551693
## spdiff.cut.fctr(-10,-1]
## 4.1547129846
## spdiff.cut.fctr(-1,0]
## 5.3348860714
## spdiff.cut.fctr(0,1]
## 4.4072812301
## spdiff.cut.fctr(1,10]
## 4.4073453219
## spdiff.cut.fctr(10,100]
## 3.8533894456
## spdiff.cut.fctr(100,1e+03]
## 1.0918155713
## sprice.d20nexp
## -1.5657841541
## sprice.log10
## -0.5769819282
## sprice.root2
## -0.1085201562
## startprice.dcm1.is9
## -0.4894970169
## startprice.dcm2.is9
## -0.0838883551
## startprice.dgt1.is9
## 0.0254181773
## startprice.dgt2.is9
## 0.1157259996
## storage.fctr16
## -0.1300313626
## storage.fctr32
## 0.0005589105
## storage.fctr64
## 0.2180851783
## storage.fctrUnknown
## 1.7336441556
## cellular.fctr0:carrier.fctrNone
## 0.1706810167
## cellular.fctr1:carrier.fctrOther
## 2.6072585524
## cellular.fctr1:carrier.fctrSprint
## 1.0270604082
## cellular.fctr1:carrier.fctrT-Mobile
## -0.2821763037
## cellular.fctr1:carrier.fctrUnknown
## -0.0179172973
## cellular.fctrUnknown:carrier.fctrUnknown
## -0.7883348835
## cellular.fctr1:carrier.fctrVerizon
## 0.3156224380
## prdl.my.fctrUnknown:.clusterid.fctr2
## 0.8932344300
## prdl.my.fctriPad1:.clusterid.fctr2
## -0.7404516345
## prdl.my.fctriPad3:.clusterid.fctr2
## 0.2701330403
## prdl.my.fctriPad4:.clusterid.fctr2
## -1.4552193885
## prdl.my.fctriPadAir:.clusterid.fctr2
## -0.3279730199
## prdl.my.fctriPadmini2:.clusterid.fctr2
## -0.6287767988
## prdl.my.fctrUnknown:.clusterid.fctr3
## -0.7598398151
## prdl.my.fctriPad1:.clusterid.fctr3
## -1.1337789530
## prdl.my.fctriPad4:.clusterid.fctr3
## -5.9466674332
## [1] "max lambda < lambdaOpt:"
## [1] "Feats mismatch between coefs_left & rght:"
## [1] "(Intercept)"
## [2] ".rnorm"
## [3] "D.chrs.n.log"
## [4] "D.chrs.pnct11.n.log"
## [5] "D.chrs.pnct13.n.log"
## [6] "D.chrs.uppr.n.log"
## [7] "D.ratio.weight.sum.wrds.n"
## [8] "D.ratio.wrds.stop.n.wrds.n"
## [9] "D.terms.post.stem.n.log"
## [10] "D.terms.post.stop.n.log"
## [11] "D.weight.post.stem.sum"
## [12] "D.weight.post.stop.sum"
## [13] "D.weight.sum"
## [14] "D.weight.sum.stem.stop.Ratio"
## [15] "D.wrds.n.log"
## [16] "D.wrds.stop.n.log"
## [17] "D.wrds.unq.n.log"
## [18] "biddable"
## [19] "cellular.fctr1"
## [20] "cellular.fctrUnknown"
## [21] "color.fctrGold"
## [22] "color.fctrSpace Gray"
## [23] "color.fctrUnknown"
## [24] "color.fctrWhite"
## [25] "condition.fctrFor parts or not working"
## [26] "condition.fctrManufacturer refurbished"
## [27] "condition.fctrNew"
## [28] "condition.fctrNew other (see details)"
## [29] "condition.fctrSeller refurbished"
## [30] "prdl.my.fctriPad1"
## [31] "prdl.my.fctriPad2"
## [32] "prdl.my.fctriPad3"
## [33] "prdl.my.fctriPad4"
## [34] "prdl.my.fctriPadAir"
## [35] "prdl.my.fctriPadAir2"
## [36] "prdl.my.fctriPadmini"
## [37] "prdl.my.fctriPadmini2"
## [38] "prdl.my.fctriPadmini3"
## [39] "spdiff.cut.fctr(-100,-10]"
## [40] "spdiff.cut.fctr(-10,-1]"
## [41] "spdiff.cut.fctr(-1,0]"
## [42] "spdiff.cut.fctr(0,1]"
## [43] "spdiff.cut.fctr(1,10]"
## [44] "spdiff.cut.fctr(10,100]"
## [45] "spdiff.cut.fctr(100,1e+03]"
## [46] "sprice.d20nexp"
## [47] "sprice.log10"
## [48] "sprice.root2"
## [49] "startprice.dcm1.is9"
## [50] "startprice.dcm2.is9"
## [51] "startprice.dgt1.is9"
## [52] "startprice.dgt2.is9"
## [53] "storage.fctr16"
## [54] "storage.fctr32"
## [55] "storage.fctr64"
## [56] "storage.fctrUnknown"
## [57] "cellular.fctr0:carrier.fctrNone"
## [58] "cellular.fctr1:carrier.fctrNone"
## [59] "cellular.fctrUnknown:carrier.fctrNone"
## [60] "cellular.fctr0:carrier.fctrOther"
## [61] "cellular.fctr1:carrier.fctrOther"
## [62] "cellular.fctrUnknown:carrier.fctrOther"
## [63] "cellular.fctr0:carrier.fctrSprint"
## [64] "cellular.fctr1:carrier.fctrSprint"
## [65] "cellular.fctrUnknown:carrier.fctrSprint"
## [66] "cellular.fctr0:carrier.fctrT-Mobile"
## [67] "cellular.fctr1:carrier.fctrT-Mobile"
## [68] "cellular.fctrUnknown:carrier.fctrT-Mobile"
## [69] "cellular.fctr0:carrier.fctrUnknown"
## [70] "cellular.fctr1:carrier.fctrUnknown"
## [71] "cellular.fctrUnknown:carrier.fctrUnknown"
## [72] "cellular.fctr0:carrier.fctrVerizon"
## [73] "cellular.fctr1:carrier.fctrVerizon"
## [74] "cellular.fctrUnknown:carrier.fctrVerizon"
## [75] "prdl.my.fctrUnknown:.clusterid.fctr2"
## [76] "prdl.my.fctriPad1:.clusterid.fctr2"
## [77] "prdl.my.fctriPad2:.clusterid.fctr2"
## [78] "prdl.my.fctriPad3:.clusterid.fctr2"
## [79] "prdl.my.fctriPad4:.clusterid.fctr2"
## [80] "prdl.my.fctriPadAir:.clusterid.fctr2"
## [81] "prdl.my.fctriPadAir2:.clusterid.fctr2"
## [82] "prdl.my.fctriPadmini:.clusterid.fctr2"
## [83] "prdl.my.fctriPadmini2:.clusterid.fctr2"
## [84] "prdl.my.fctriPadmini3:.clusterid.fctr2"
## [85] "prdl.my.fctrUnknown:.clusterid.fctr3"
## [86] "prdl.my.fctriPad1:.clusterid.fctr3"
## [87] "prdl.my.fctriPad2:.clusterid.fctr3"
## [88] "prdl.my.fctriPad3:.clusterid.fctr3"
## [89] "prdl.my.fctriPad4:.clusterid.fctr3"
## [90] "prdl.my.fctriPadAir:.clusterid.fctr3"
## [91] "prdl.my.fctriPadAir2:.clusterid.fctr3"
## [92] "prdl.my.fctriPadmini:.clusterid.fctr3"
## [93] "prdl.my.fctriPadmini2:.clusterid.fctr3"
## [94] "prdl.my.fctriPadmini3:.clusterid.fctr3"
## sold.fctr sold.fctr.predict.RFE.X.glmnet.N
## 1 N 467
## 2 Y 63
## sold.fctr.predict.RFE.X.glmnet.Y
## 1 81
## 2 409
## Prediction
## Reference N Y
## N 467 81
## Y 63 409
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 8.588235e-01 7.168196e-01 8.359295e-01 8.796230e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 2.677951e-106 1.565804e-01
## sold.fctr sold.fctr.predict.RFE.X.glmnet.N
## 1 N 309
## 2 Y 60
## sold.fctr.predict.RFE.X.glmnet.Y
## 1 138
## 2 323
## Prediction
## Reference N Y
## N 309 138
## Y 60 323
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.614458e-01 5.269348e-01 7.309494e-01 7.900724e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 2.413828e-40 4.446039e-08
## id
## 1 RFE.X.glmnet
## feats
## 1 sprice.d20nexp,sprice.log10,sprice.root2,biddable,spdiff.cut.fctr,cellular.fctr,prdl.my.fctr,D.ratio.wrds.stop.n.wrds.n,storage.fctr,condition.fctr,D.weight.sum.stem.stop.Ratio,color.fctr,startprice.dcm2.is9,.rnorm,D.chrs.pnct11.n.log,D.wrds.stop.n.log,startprice.dcm1.is9,startprice.dgt2.is9,D.ratio.weight.sum.wrds.n,D.chrs.pnct13.n.log,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,D.wrds.n.log,D.chrs.uppr.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,D.terms.post.stem.n.log,D.chrs.n.log,startprice.dgt1.is9,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 25 15.219 0.534
## max.AUCpROC.fit max.Sens.fit max.Spec.fit max.AUCROCR.fit
## 1 0.8626593 0.9032847 0.8220339 0.9355785
## opt.prob.threshold.fit max.f.score.fit max.Accuracy.fit
## 1 0.4 0.8503119 0.8303886
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.8359295 0.879623 0.6576319
## max.AUCpROC.OOB max.Sens.OOB max.Spec.OOB max.AUCROCR.OOB
## 1 0.7835089 0.8255034 0.7415144 0.8721444
## opt.prob.threshold.OOB max.f.score.OOB max.Accuracy.OOB
## 1 0.3 0.7654028 0.7614458
## max.AccuracyLower.OOB max.AccuracyUpper.OOB max.Kappa.OOB
## 1 0.7309494 0.7900724 0.5269348
## max.AccuracySD.fit max.KappaSD.fit
## 1 0.01126649 0.02276618
## label step_major step_minor label_minor bgn end
## 3 fit.models_1_RFE.X 2 1 glmnet 424.011 443.927
## 4 fit.models_1_RFE.X 2 2 glm 443.928 NA
## elapsed
## 3 19.916
## 4 NA
## [1] "fitting model: RFE.X.glm"
## [1] " indep_vars: sprice.d20nexp,sprice.log10,sprice.root2,biddable,spdiff.cut.fctr,cellular.fctr,prdl.my.fctr,D.ratio.wrds.stop.n.wrds.n,storage.fctr,condition.fctr,D.weight.sum.stem.stop.Ratio,color.fctr,startprice.dcm2.is9,.rnorm,D.chrs.pnct11.n.log,D.wrds.stop.n.log,startprice.dcm1.is9,startprice.dgt2.is9,D.ratio.weight.sum.wrds.n,D.chrs.pnct13.n.log,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,D.wrds.n.log,D.chrs.uppr.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,D.terms.post.stem.n.log,D.chrs.n.log,startprice.dgt1.is9,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr"
## + Fold1.Rep1: parameter=none## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold1.Rep1: parameter=none
## + Fold2.Rep1: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold2.Rep1: parameter=none
## + Fold3.Rep1: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold3.Rep1: parameter=none
## + Fold1.Rep2: parameter=none## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold1.Rep2: parameter=none
## + Fold2.Rep2: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold2.Rep2: parameter=none
## + Fold3.Rep2: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold3.Rep2: parameter=none
## + Fold1.Rep3: parameter=none## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold1.Rep3: parameter=none
## + Fold2.Rep3: parameter=none## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold2.Rep3: parameter=none
## + Fold3.Rep3: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold3.Rep3: parameter=none
## Aggregating results
## Fitting final model on full training set## Warning: not plotting observations with leverage one:
## 683
## Warning: not plotting observations with leverage one:
## 683
##
## Call:
## NULL
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.8799 -0.4583 -0.0814 0.3151 3.4426
##
## Coefficients: (26 not defined because of singularities)
## Estimate Std. Error z value
## (Intercept) -2.401e+01 3.068e+01 -0.783
## .rnorm -1.788e-02 1.057e-01 -0.169
## D.chrs.n.log -6.596e+00 4.632e+00 -1.424
## D.chrs.pnct11.n.log 6.989e-01 4.210e-01 1.660
## D.chrs.pnct13.n.log 1.283e-01 4.155e-01 0.309
## D.chrs.uppr.n.log 4.347e+00 4.277e+00 1.016
## D.ratio.weight.sum.wrds.n 2.759e-01 3.276e-01 0.842
## D.ratio.wrds.stop.n.wrds.n -1.040e+01 4.111e+00 -2.530
## D.terms.post.stem.n.log 2.354e+01 1.124e+01 2.094
## D.terms.post.stop.n.log -3.028e+01 1.119e+01 -2.706
## D.weight.post.stem.sum -7.805e+00 5.538e+00 -1.409
## D.weight.post.stop.sum 7.281e+00 5.343e+00 1.363
## D.weight.sum NA NA NA
## D.weight.sum.stem.stop.Ratio 3.529e+01 3.102e+01 1.138
## D.wrds.n.log 7.889e+00 2.575e+00 3.064
## D.wrds.stop.n.log 2.840e-01 7.047e-01 0.403
## D.wrds.unq.n.log NA NA NA
## biddable 2.630e+00 2.970e-01 8.856
## cellular.fctr1 -2.086e-01 3.623e-01 -0.576
## cellular.fctrUnknown -1.826e+00 6.228e-01 -2.931
## color.fctrGold 9.378e-02 7.237e-01 0.130
## `color.fctrSpace Gray` -3.219e-01 4.232e-01 -0.761
## color.fctrUnknown 3.371e-01 2.968e-01 1.136
## color.fctrWhite -1.613e-01 3.303e-01 -0.488
## `condition.fctrFor parts or not working` -9.316e-01 4.578e-01 -2.035
## `condition.fctrManufacturer refurbished` -1.304e+00 8.416e-01 -1.549
## condition.fctrNew 5.179e-02 4.037e-01 0.128
## `condition.fctrNew other (see details)` 6.535e-01 5.671e-01 1.152
## `condition.fctrSeller refurbished` -1.092e+00 4.754e-01 -2.296
## prdl.my.fctriPad1 -6.632e-01 6.929e-01 -0.957
## prdl.my.fctriPad2 -4.322e-01 6.111e-01 -0.707
## prdl.my.fctriPad3 2.127e-01 6.437e-01 0.330
## prdl.my.fctriPad4 1.449e+00 7.149e-01 2.027
## prdl.my.fctriPadAir 1.646e+00 7.081e-01 2.324
## prdl.my.fctriPadAir2 1.802e+00 7.860e-01 2.293
## prdl.my.fctriPadmini -6.314e-01 5.989e-01 -1.054
## prdl.my.fctriPadmini2 1.177e+00 7.368e-01 1.597
## prdl.my.fctriPadmini3 5.642e-01 8.039e-01 0.702
## `spdiff.cut.fctr(-100,-10]` 2.478e+00 4.231e-01 5.857
## `spdiff.cut.fctr(-10,-1]` 4.362e+00 5.586e-01 7.809
## `spdiff.cut.fctr(-1,0]` 5.453e+00 1.021e+00 5.339
## `spdiff.cut.fctr(0,1]` 4.625e+00 1.319e+00 3.508
## `spdiff.cut.fctr(1,10]` 4.774e+00 5.632e-01 8.476
## `spdiff.cut.fctr(10,100]` 4.088e+00 5.065e-01 8.071
## `spdiff.cut.fctr(100,1e+03]` 1.217e+00 8.025e-01 1.517
## sprice.d20nexp -2.326e+00 3.455e+00 -0.673
## sprice.log10 -8.826e-01 1.088e+00 -0.811
## sprice.root2 -5.536e-02 1.602e-01 -0.345
## startprice.dcm1.is9 -5.127e-01 4.195e-01 -1.222
## startprice.dcm2.is9 -5.052e-02 4.086e-01 -0.124
## startprice.dgt1.is9 -1.686e-02 2.977e-01 -0.057
## startprice.dgt2.is9 6.216e-02 3.276e-01 0.190
## storage.fctr16 -8.848e-03 7.434e-01 -0.012
## storage.fctr32 1.539e-01 7.550e-01 0.204
## storage.fctr64 2.955e-01 7.377e-01 0.401
## storage.fctrUnknown 1.900e+00 9.548e-01 1.990
## `cellular.fctr0:carrier.fctrNone` NA NA NA
## `cellular.fctr1:carrier.fctrNone` NA NA NA
## `cellular.fctrUnknown:carrier.fctrNone` NA NA NA
## `cellular.fctr0:carrier.fctrOther` NA NA NA
## `cellular.fctr1:carrier.fctrOther` 1.273e+01 2.400e+03 0.005
## `cellular.fctrUnknown:carrier.fctrOther` NA NA NA
## `cellular.fctr0:carrier.fctrSprint` NA NA NA
## `cellular.fctr1:carrier.fctrSprint` 9.312e-01 7.822e-01 1.190
## `cellular.fctrUnknown:carrier.fctrSprint` NA NA NA
## `cellular.fctr0:carrier.fctrT-Mobile` NA NA NA
## `cellular.fctr1:carrier.fctrT-Mobile` 2.874e-02 1.283e+00 0.022
## `cellular.fctrUnknown:carrier.fctrT-Mobile` NA NA NA
## `cellular.fctr0:carrier.fctrUnknown` NA NA NA
## `cellular.fctr1:carrier.fctrUnknown` -6.309e-03 6.176e-01 -0.010
## `cellular.fctrUnknown:carrier.fctrUnknown` NA NA NA
## `cellular.fctr0:carrier.fctrVerizon` NA NA NA
## `cellular.fctr1:carrier.fctrVerizon` 4.164e-01 5.119e-01 0.813
## `cellular.fctrUnknown:carrier.fctrVerizon` NA NA NA
## `prdl.my.fctrUnknown:.clusterid.fctr2` 6.062e-01 8.836e-01 0.686
## `prdl.my.fctriPad1:.clusterid.fctr2` -1.344e+00 8.787e-01 -1.529
## `prdl.my.fctriPad2:.clusterid.fctr2` NA NA NA
## `prdl.my.fctriPad3:.clusterid.fctr2` 6.740e-02 9.757e-01 0.069
## `prdl.my.fctriPad4:.clusterid.fctr2` -1.954e+00 1.355e+00 -1.442
## `prdl.my.fctriPadAir:.clusterid.fctr2` -3.340e-01 9.500e-01 -0.352
## `prdl.my.fctriPadAir2:.clusterid.fctr2` NA NA NA
## `prdl.my.fctriPadmini:.clusterid.fctr2` NA NA NA
## `prdl.my.fctriPadmini2:.clusterid.fctr2` -6.927e-01 1.494e+00 -0.464
## `prdl.my.fctriPadmini3:.clusterid.fctr2` NA NA NA
## `prdl.my.fctrUnknown:.clusterid.fctr3` -7.640e-02 1.280e+00 -0.060
## `prdl.my.fctriPad1:.clusterid.fctr3` -1.353e+00 8.714e-01 -1.553
## `prdl.my.fctriPad2:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPad3:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPad4:.clusterid.fctr3` -1.652e+01 4.931e+02 -0.034
## `prdl.my.fctriPadAir:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPadAir2:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPadmini:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPadmini2:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPadmini3:.clusterid.fctr3` NA NA NA
## Pr(>|z|)
## (Intercept) 0.433733
## .rnorm 0.865713
## D.chrs.n.log 0.154373
## D.chrs.pnct11.n.log 0.096884 .
## D.chrs.pnct13.n.log 0.757516
## D.chrs.uppr.n.log 0.309512
## D.ratio.weight.sum.wrds.n 0.399691
## D.ratio.wrds.stop.n.wrds.n 0.011396 *
## D.terms.post.stem.n.log 0.036258 *
## D.terms.post.stop.n.log 0.006800 **
## D.weight.post.stem.sum 0.158704
## D.weight.post.stop.sum 0.172972
## D.weight.sum NA
## D.weight.sum.stem.stop.Ratio 0.255188
## D.wrds.n.log 0.002185 **
## D.wrds.stop.n.log 0.686961
## D.wrds.unq.n.log NA
## biddable < 2e-16 ***
## cellular.fctr1 0.564862
## cellular.fctrUnknown 0.003376 **
## color.fctrGold 0.896906
## `color.fctrSpace Gray` 0.446886
## color.fctrUnknown 0.256140
## color.fctrWhite 0.625236
## `condition.fctrFor parts or not working` 0.041845 *
## `condition.fctrManufacturer refurbished` 0.121369
## condition.fctrNew 0.897934
## `condition.fctrNew other (see details)` 0.249218
## `condition.fctrSeller refurbished` 0.021667 *
## prdl.my.fctriPad1 0.338494
## prdl.my.fctriPad2 0.479456
## prdl.my.fctriPad3 0.741110
## prdl.my.fctriPad4 0.042655 *
## prdl.my.fctriPadAir 0.020109 *
## prdl.my.fctriPadAir2 0.021849 *
## prdl.my.fctriPadmini 0.291796
## prdl.my.fctriPadmini2 0.110192
## prdl.my.fctriPadmini3 0.482795
## `spdiff.cut.fctr(-100,-10]` 4.71e-09 ***
## `spdiff.cut.fctr(-10,-1]` 5.76e-15 ***
## `spdiff.cut.fctr(-1,0]` 9.34e-08 ***
## `spdiff.cut.fctr(0,1]` 0.000452 ***
## `spdiff.cut.fctr(1,10]` < 2e-16 ***
## `spdiff.cut.fctr(10,100]` 6.99e-16 ***
## `spdiff.cut.fctr(100,1e+03]` 0.129375
## sprice.d20nexp 0.500735
## sprice.log10 0.417449
## sprice.root2 0.729739
## startprice.dcm1.is9 0.221611
## startprice.dcm2.is9 0.901608
## startprice.dgt1.is9 0.954845
## startprice.dgt2.is9 0.849498
## storage.fctr16 0.990505
## storage.fctr32 0.838496
## storage.fctr64 0.688748
## storage.fctrUnknown 0.046572 *
## `cellular.fctr0:carrier.fctrNone` NA
## `cellular.fctr1:carrier.fctrNone` NA
## `cellular.fctrUnknown:carrier.fctrNone` NA
## `cellular.fctr0:carrier.fctrOther` NA
## `cellular.fctr1:carrier.fctrOther` 0.995769
## `cellular.fctrUnknown:carrier.fctrOther` NA
## `cellular.fctr0:carrier.fctrSprint` NA
## `cellular.fctr1:carrier.fctrSprint` 0.233863
## `cellular.fctrUnknown:carrier.fctrSprint` NA
## `cellular.fctr0:carrier.fctrT-Mobile` NA
## `cellular.fctr1:carrier.fctrT-Mobile` 0.982129
## `cellular.fctrUnknown:carrier.fctrT-Mobile` NA
## `cellular.fctr0:carrier.fctrUnknown` NA
## `cellular.fctr1:carrier.fctrUnknown` 0.991849
## `cellular.fctrUnknown:carrier.fctrUnknown` NA
## `cellular.fctr0:carrier.fctrVerizon` NA
## `cellular.fctr1:carrier.fctrVerizon` 0.416013
## `cellular.fctrUnknown:carrier.fctrVerizon` NA
## `prdl.my.fctrUnknown:.clusterid.fctr2` 0.492623
## `prdl.my.fctriPad1:.clusterid.fctr2` 0.126225
## `prdl.my.fctriPad2:.clusterid.fctr2` NA
## `prdl.my.fctriPad3:.clusterid.fctr2` 0.944930
## `prdl.my.fctriPad4:.clusterid.fctr2` 0.149436
## `prdl.my.fctriPadAir:.clusterid.fctr2` 0.725134
## `prdl.my.fctriPadAir2:.clusterid.fctr2` NA
## `prdl.my.fctriPadmini:.clusterid.fctr2` NA
## `prdl.my.fctriPadmini2:.clusterid.fctr2` 0.642927
## `prdl.my.fctriPadmini3:.clusterid.fctr2` NA
## `prdl.my.fctrUnknown:.clusterid.fctr3` 0.952415
## `prdl.my.fctriPad1:.clusterid.fctr3` 0.120433
## `prdl.my.fctriPad2:.clusterid.fctr3` NA
## `prdl.my.fctriPad3:.clusterid.fctr3` NA
## `prdl.my.fctriPad4:.clusterid.fctr3` 0.973270
## `prdl.my.fctriPadAir:.clusterid.fctr3` NA
## `prdl.my.fctriPadAir2:.clusterid.fctr3` NA
## `prdl.my.fctriPadmini:.clusterid.fctr3` NA
## `prdl.my.fctriPadmini2:.clusterid.fctr3` NA
## `prdl.my.fctriPadmini3:.clusterid.fctr3` NA
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 1408.35 on 1019 degrees of freedom
## Residual deviance: 633.39 on 952 degrees of freedom
## AIC: 769.39
##
## Number of Fisher Scoring iterations: 15## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading
## sold.fctr sold.fctr.predict.RFE.X.glm.N sold.fctr.predict.RFE.X.glm.Y
## 1 N 469 79
## 2 Y 65 407
## Prediction
## Reference N Y
## N 469 79
## Y 65 407
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 8.588235e-01 7.166535e-01 8.359295e-01 8.796230e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 2.677951e-106 2.786605e-01## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading
## sold.fctr sold.fctr.predict.RFE.X.glm.N sold.fctr.predict.RFE.X.glm.Y
## 1 N 367 80
## 2 Y 96 287
## Prediction
## Reference N Y
## N 367 80
## Y 96 287
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.879518e-01 5.720872e-01 7.585420e-01 8.152999e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 1.092681e-50 2.581950e-01
## id
## 1 RFE.X.glm
## feats
## 1 sprice.d20nexp,sprice.log10,sprice.root2,biddable,spdiff.cut.fctr,cellular.fctr,prdl.my.fctr,D.ratio.wrds.stop.n.wrds.n,storage.fctr,condition.fctr,D.weight.sum.stem.stop.Ratio,color.fctr,startprice.dcm2.is9,.rnorm,D.chrs.pnct11.n.log,D.wrds.stop.n.log,startprice.dcm1.is9,startprice.dgt2.is9,D.ratio.weight.sum.wrds.n,D.chrs.pnct13.n.log,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,D.wrds.n.log,D.chrs.uppr.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,D.terms.post.stem.n.log,D.chrs.n.log,startprice.dgt1.is9,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 1 2.89 0.192
## max.AUCpROC.fit max.Sens.fit max.Spec.fit max.AUCROCR.fit
## 1 0.8597751 0.899635 0.8199153 0.9392823
## opt.prob.threshold.fit max.f.score.fit max.Accuracy.fit
## 1 0.4 0.8496868 0.8307135
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.8359295 0.879623 0.6585198
## max.AUCpROC.OOB max.Sens.OOB max.Spec.OOB max.AUCROCR.OOB
## 1 0.7851882 0.8210291 0.7493473 0.8700475
## opt.prob.threshold.OOB max.f.score.OOB max.Accuracy.OOB
## 1 0.5 0.7653333 0.7879518
## max.AccuracyLower.OOB max.AccuracyUpper.OOB max.Kappa.OOB
## 1 0.758542 0.8152999 0.5720872
## max.AccuracySD.fit max.KappaSD.fit
## 1 0.01447782 0.02900194
## label step_major step_minor label_minor bgn end
## 4 fit.models_1_RFE.X 2 2 glm 443.928 451.375
## 5 fit.models_1_All.X 3 0 setup 451.376 NA
## elapsed
## 4 7.447
## 5 NA
## label step_major step_minor label_minor bgn end
## 5 fit.models_1_All.X 3 0 setup 451.376 451.383
## 6 fit.models_1_All.X 3 1 glmnet 451.384 NA
## elapsed
## 5 0.007
## 6 NA
## [1] "fitting model: All.X.glmnet"
## [1] " indep_vars: biddable,sprice.d20nexp,spdiff.cut.fctr,D.ratio.wrds.stop.n.wrds.n,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,startprice.dcm1.is9,storage.fctr,D.chrs.pnct11.n.log,D.chrs.pnct13.n.log,startprice.dgt2.is9,color.fctr,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,prdl.my.fctr,D.chrs.uppr.n.log,D.chrs.n.log,D.wrds.n.log,D.terms.post.stem.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.log10,sprice.root2,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0.1, lambda = 0.000267 on full training set## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst
## = list(id.prefix = mdl_id_pfx, : model's bestTune found at an extreme of
## tuneGrid for parameter: alpha## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst
## = list(id.prefix = mdl_id_pfx, : model's bestTune found at an extreme of
## tuneGrid for parameter: lambda
## Length Class Mode
## a0 100 -none- numeric
## beta 9300 dgCMatrix S4
## df 100 -none- numeric
## dim 2 -none- numeric
## lambda 100 -none- numeric
## dev.ratio 100 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 93 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept)
## 5.9298599950
## .rnorm
## -0.0367487387
## D.chrs.n.log
## -0.2852958123
## D.chrs.pnct11.n.log
## 0.4714450805
## D.chrs.pnct13.n.log
## -0.1653784579
## D.chrs.uppr.n.log
## -0.0116026565
## D.ratio.weight.sum.wrds.n
## -0.0593199934
## D.ratio.wrds.stop.n.wrds.n
## -2.9071388332
## D.terms.post.stem.n.log
## -0.2864766681
## D.terms.post.stop.n.log
## -0.9265313808
## D.weight.post.stem.sum
## -0.1209650771
## D.weight.sum
## -0.1180951219
## D.weight.sum.stem.stop.Ratio
## -2.9013605803
## D.wrds.n.log
## 1.2607015439
## D.wrds.stop.n.log
## 0.1152103538
## D.wrds.unq.n.log
## -0.1770678274
## biddable
## 2.5343114368
## cellular.fctrUnknown
## -0.7922522457
## color.fctrGold
## 0.0871372337
## color.fctrSpace Gray
## -0.3474018561
## color.fctrUnknown
## 0.3560862355
## color.fctrWhite
## -0.2253657549
## condition.fctrFor parts or not working
## -0.7365978083
## condition.fctrManufacturer refurbished
## -1.1969859100
## condition.fctrNew
## 0.0297417566
## condition.fctrNew other (see details)
## 0.7229925582
## condition.fctrSeller refurbished
## -1.0271638293
## prdl.my.fctriPad1
## -0.6856562063
## prdl.my.fctriPad2
## -0.4622200902
## prdl.my.fctriPad3
## 0.1837795731
## prdl.my.fctriPad4
## 1.4053112779
## prdl.my.fctriPadAir
## 1.5486082406
## prdl.my.fctriPadAir2
## 1.9064724435
## prdl.my.fctriPadmini
## -0.4755225796
## prdl.my.fctriPadmini2
## 1.3324979842
## prdl.my.fctriPadmini3
## 0.7573021470
## spdiff.cut.fctr(-100,-10]
## 2.3577551693
## spdiff.cut.fctr(-10,-1]
## 4.1547129846
## spdiff.cut.fctr(-1,0]
## 5.3348860714
## spdiff.cut.fctr(0,1]
## 4.4072812301
## spdiff.cut.fctr(1,10]
## 4.4073453219
## spdiff.cut.fctr(10,100]
## 3.8533894456
## spdiff.cut.fctr(100,1e+03]
## 1.0918155713
## sprice.d20nexp
## -1.5657841541
## sprice.log10
## -0.5769819282
## sprice.root2
## -0.1085201562
## startprice.dcm1.is9
## -0.4894970169
## startprice.dcm2.is9
## -0.0838883551
## startprice.dgt1.is9
## 0.0254181773
## startprice.dgt2.is9
## 0.1157259996
## storage.fctr16
## -0.1300313626
## storage.fctr32
## 0.0005589105
## storage.fctr64
## 0.2180851783
## storage.fctrUnknown
## 1.7336441556
## cellular.fctr0:carrier.fctrNone
## 0.1706810167
## cellular.fctr1:carrier.fctrOther
## 2.6072585524
## cellular.fctr1:carrier.fctrSprint
## 1.0270604082
## cellular.fctr1:carrier.fctrT-Mobile
## -0.2821763037
## cellular.fctr1:carrier.fctrUnknown
## -0.0179172973
## cellular.fctrUnknown:carrier.fctrUnknown
## -0.7883348835
## cellular.fctr1:carrier.fctrVerizon
## 0.3156224380
## prdl.my.fctrUnknown:.clusterid.fctr2
## 0.8932344300
## prdl.my.fctriPad1:.clusterid.fctr2
## -0.7404516345
## prdl.my.fctriPad3:.clusterid.fctr2
## 0.2701330403
## prdl.my.fctriPad4:.clusterid.fctr2
## -1.4552193885
## prdl.my.fctriPadAir:.clusterid.fctr2
## -0.3279730199
## prdl.my.fctriPadmini2:.clusterid.fctr2
## -0.6287767988
## prdl.my.fctrUnknown:.clusterid.fctr3
## -0.7598398151
## prdl.my.fctriPad1:.clusterid.fctr3
## -1.1337789530
## prdl.my.fctriPad4:.clusterid.fctr3
## -5.9466674332
## [1] "max lambda < lambdaOpt:"
## [1] "Feats mismatch between coefs_left & rght:"
## [1] "(Intercept)"
## [2] ".rnorm"
## [3] "D.chrs.n.log"
## [4] "D.chrs.pnct11.n.log"
## [5] "D.chrs.pnct13.n.log"
## [6] "D.chrs.uppr.n.log"
## [7] "D.ratio.weight.sum.wrds.n"
## [8] "D.ratio.wrds.stop.n.wrds.n"
## [9] "D.terms.post.stem.n.log"
## [10] "D.terms.post.stop.n.log"
## [11] "D.weight.post.stem.sum"
## [12] "D.weight.post.stop.sum"
## [13] "D.weight.sum"
## [14] "D.weight.sum.stem.stop.Ratio"
## [15] "D.wrds.n.log"
## [16] "D.wrds.stop.n.log"
## [17] "D.wrds.unq.n.log"
## [18] "biddable"
## [19] "cellular.fctr1"
## [20] "cellular.fctrUnknown"
## [21] "color.fctrGold"
## [22] "color.fctrSpace Gray"
## [23] "color.fctrUnknown"
## [24] "color.fctrWhite"
## [25] "condition.fctrFor parts or not working"
## [26] "condition.fctrManufacturer refurbished"
## [27] "condition.fctrNew"
## [28] "condition.fctrNew other (see details)"
## [29] "condition.fctrSeller refurbished"
## [30] "prdl.my.fctriPad1"
## [31] "prdl.my.fctriPad2"
## [32] "prdl.my.fctriPad3"
## [33] "prdl.my.fctriPad4"
## [34] "prdl.my.fctriPadAir"
## [35] "prdl.my.fctriPadAir2"
## [36] "prdl.my.fctriPadmini"
## [37] "prdl.my.fctriPadmini2"
## [38] "prdl.my.fctriPadmini3"
## [39] "spdiff.cut.fctr(-100,-10]"
## [40] "spdiff.cut.fctr(-10,-1]"
## [41] "spdiff.cut.fctr(-1,0]"
## [42] "spdiff.cut.fctr(0,1]"
## [43] "spdiff.cut.fctr(1,10]"
## [44] "spdiff.cut.fctr(10,100]"
## [45] "spdiff.cut.fctr(100,1e+03]"
## [46] "sprice.d20nexp"
## [47] "sprice.log10"
## [48] "sprice.root2"
## [49] "startprice.dcm1.is9"
## [50] "startprice.dcm2.is9"
## [51] "startprice.dgt1.is9"
## [52] "startprice.dgt2.is9"
## [53] "storage.fctr16"
## [54] "storage.fctr32"
## [55] "storage.fctr64"
## [56] "storage.fctrUnknown"
## [57] "cellular.fctr0:carrier.fctrNone"
## [58] "cellular.fctr1:carrier.fctrNone"
## [59] "cellular.fctrUnknown:carrier.fctrNone"
## [60] "cellular.fctr0:carrier.fctrOther"
## [61] "cellular.fctr1:carrier.fctrOther"
## [62] "cellular.fctrUnknown:carrier.fctrOther"
## [63] "cellular.fctr0:carrier.fctrSprint"
## [64] "cellular.fctr1:carrier.fctrSprint"
## [65] "cellular.fctrUnknown:carrier.fctrSprint"
## [66] "cellular.fctr0:carrier.fctrT-Mobile"
## [67] "cellular.fctr1:carrier.fctrT-Mobile"
## [68] "cellular.fctrUnknown:carrier.fctrT-Mobile"
## [69] "cellular.fctr0:carrier.fctrUnknown"
## [70] "cellular.fctr1:carrier.fctrUnknown"
## [71] "cellular.fctrUnknown:carrier.fctrUnknown"
## [72] "cellular.fctr0:carrier.fctrVerizon"
## [73] "cellular.fctr1:carrier.fctrVerizon"
## [74] "cellular.fctrUnknown:carrier.fctrVerizon"
## [75] "prdl.my.fctrUnknown:.clusterid.fctr2"
## [76] "prdl.my.fctriPad1:.clusterid.fctr2"
## [77] "prdl.my.fctriPad2:.clusterid.fctr2"
## [78] "prdl.my.fctriPad3:.clusterid.fctr2"
## [79] "prdl.my.fctriPad4:.clusterid.fctr2"
## [80] "prdl.my.fctriPadAir:.clusterid.fctr2"
## [81] "prdl.my.fctriPadAir2:.clusterid.fctr2"
## [82] "prdl.my.fctriPadmini:.clusterid.fctr2"
## [83] "prdl.my.fctriPadmini2:.clusterid.fctr2"
## [84] "prdl.my.fctriPadmini3:.clusterid.fctr2"
## [85] "prdl.my.fctrUnknown:.clusterid.fctr3"
## [86] "prdl.my.fctriPad1:.clusterid.fctr3"
## [87] "prdl.my.fctriPad2:.clusterid.fctr3"
## [88] "prdl.my.fctriPad3:.clusterid.fctr3"
## [89] "prdl.my.fctriPad4:.clusterid.fctr3"
## [90] "prdl.my.fctriPadAir:.clusterid.fctr3"
## [91] "prdl.my.fctriPadAir2:.clusterid.fctr3"
## [92] "prdl.my.fctriPadmini:.clusterid.fctr3"
## [93] "prdl.my.fctriPadmini2:.clusterid.fctr3"
## [94] "prdl.my.fctriPadmini3:.clusterid.fctr3"
## sold.fctr sold.fctr.predict.All.X.glmnet.N
## 1 N 467
## 2 Y 63
## sold.fctr.predict.All.X.glmnet.Y
## 1 81
## 2 409
## Prediction
## Reference N Y
## N 467 81
## Y 63 409
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 8.588235e-01 7.168196e-01 8.359295e-01 8.796230e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 2.677951e-106 1.565804e-01
## sold.fctr sold.fctr.predict.All.X.glmnet.N
## 1 N 309
## 2 Y 60
## sold.fctr.predict.All.X.glmnet.Y
## 1 138
## 2 323
## Prediction
## Reference N Y
## N 309 138
## Y 60 323
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.614458e-01 5.269348e-01 7.309494e-01 7.900724e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 2.413828e-40 4.446039e-08
## id
## 1 All.X.glmnet
## feats
## 1 biddable,sprice.d20nexp,spdiff.cut.fctr,D.ratio.wrds.stop.n.wrds.n,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,startprice.dcm1.is9,storage.fctr,D.chrs.pnct11.n.log,D.chrs.pnct13.n.log,startprice.dgt2.is9,color.fctr,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,prdl.my.fctr,D.chrs.uppr.n.log,D.chrs.n.log,D.wrds.n.log,D.terms.post.stem.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.log10,sprice.root2,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 25 14.826 0.545
## max.AUCpROC.fit max.Sens.fit max.Spec.fit max.AUCROCR.fit
## 1 0.8626593 0.9032847 0.8220339 0.9355785
## opt.prob.threshold.fit max.f.score.fit max.Accuracy.fit
## 1 0.4 0.8503119 0.8303886
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.8359295 0.879623 0.6576319
## max.AUCpROC.OOB max.Sens.OOB max.Spec.OOB max.AUCROCR.OOB
## 1 0.7835089 0.8255034 0.7415144 0.8721444
## opt.prob.threshold.OOB max.f.score.OOB max.Accuracy.OOB
## 1 0.3 0.7654028 0.7614458
## max.AccuracyLower.OOB max.AccuracyUpper.OOB max.Kappa.OOB
## 1 0.7309494 0.7900724 0.5269348
## max.AccuracySD.fit max.KappaSD.fit
## 1 0.01126649 0.02276618
## label step_major step_minor label_minor bgn
## 6 fit.models_1_All.X 3 1 glmnet 451.384
## 7 fit.models_1_Best.Interact 4 0 setup 470.799
## end elapsed
## 6 470.799 19.415
## 7 NA NA
## label step_major step_minor
## 7 fit.models_1_Best.Interact 4 0
## 8 fit.models_1_Max.cor.Y.rcv.1X1.Interact 4 1
## label_minor bgn end elapsed
## 7 setup 470.799 470.84 0.041
## 8 glmnet 470.840 NA NA
## [1] "fitting model: Max.cor.Y.rcv.1X1.Interact.glmnet"
## [1] " indep_vars: sprice.root2,biddable"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0.325, lambda = 0.124 on full training set## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst
## = list(id.prefix = mdl_id_pfx, : model's bestTune found at an extreme of
## tuneGrid for parameter: lambda
## Length Class Mode
## a0 67 -none- numeric
## beta 134 dgCMatrix S4
## df 67 -none- numeric
## dim 2 -none- numeric
## lambda 67 -none- numeric
## dev.ratio 67 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 2 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.36344652 1.18295779 -0.08216503
## [1] "max lambda < lambdaOpt:"
## (Intercept) biddable sprice.root2
## 0.39118688 1.22752496 -0.08592317
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.1X1.Interact.glmnet.N
## 1 N 457
## 2 Y 118
## sold.fctr.predict.Max.cor.Y.rcv.1X1.Interact.glmnet.Y
## 1 91
## 2 354
## Prediction
## Reference N Y
## N 457 91
## Y 118 354
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.950980e-01 5.862671e-01 7.690020e-01 8.194783e-01 5.372549e-01
## AccuracyPValue McnemarPValue
## 3.331019e-66 7.210452e-02
## sold.fctr sold.fctr.predict.Max.cor.Y.rcv.1X1.Interact.glmnet.N
## 1 N 341
## 2 Y 88
## sold.fctr.predict.Max.cor.Y.rcv.1X1.Interact.glmnet.Y
## 1 106
## 2 295
## Prediction
## Reference N Y
## N 341 106
## Y 88 295
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 7.662651e-01 5.313109e-01 7.359543e-01 7.946712e-01 5.385542e-01
## AccuracyPValue McnemarPValue
## 4.058545e-42 2.222645e-01
## id feats max.nTuningRuns
## 1 Max.cor.Y.rcv.1X1.Interact.glmnet sprice.root2,biddable 25
## min.elapsedtime.everything min.elapsedtime.final max.AUCpROC.fit
## 1 2.319 0.016 0.7919708
## max.Sens.fit max.Spec.fit max.AUCROCR.fit opt.prob.threshold.fit
## 1 0.8339416 0.75 0.8622147 0.5
## max.f.score.fit max.Accuracy.fit max.AccuracyLower.fit
## 1 0.7720829 0.7954221 0.769002
## max.AccuracyUpper.fit max.Kappa.fit max.AUCpROC.OOB max.Sens.OOB
## 1 0.8194783 0.5866808 0.760384 0.8210291
## max.Spec.OOB max.AUCROCR.OOB opt.prob.threshold.OOB max.f.score.OOB
## 1 0.6997389 0.819747 0.4 0.752551
## max.Accuracy.OOB max.AccuracyLower.OOB max.AccuracyUpper.OOB
## 1 0.7662651 0.7359543 0.7946712
## max.Kappa.OOB max.AccuracySD.fit max.KappaSD.fit
## 1 0.5313109 0.0200887 0.04087803# Check if other preProcess methods improve model performance
mdl_id <- orderBy(get_model_sel_frmla(), glb_models_df)[1, "id"]
indep_vars_vctr <-
trim(unlist(strsplit(glb_models_df[glb_models_df$id == mdl_id, "feats"], "[,]")))
method <- tail(unlist(strsplit(mdl_id, "[.]")), 1)
mdl_id_pfx <- paste0(head(unlist(strsplit(mdl_id, "[.]")), -1), collapse=".")
if (!is.null(glbObsFitOutliers[[mdl_id_pfx]])) {
fitobs_df <- glb_fitobs_df[!(glb_fitobs_df[, glb_id_var] %in%
glbObsFitOutliers[[mdl_id_pfx]]), ]
} else fitobs_df <- glb_fitobs_df
for (prePr in glb_preproc_methods) {
# The operations are applied in this order:
# Box-Cox/Yeo-Johnson transformation, centering, scaling, range, imputation, PCA, ICA then spatial sign.
ret_lst <- myfit_mdl(mdl_specs_lst=myinit_mdl_specs_lst(mdl_specs_lst=list(
id.prefix=mdl_id_pfx,
type=glb_model_type, tune.df=glb_tune_models_df,
trainControl.method="repeatedcv",
trainControl.number=glb_rcv_n_folds,
trainControl.repeats=glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method=method, train.preProcess=prePr)),
indep_vars=indep_vars_vctr, rsp_var=glb_rsp_var,
fit_df=fitobs_df, OOB_df=glb_OOBobs_df)
}
# If (All|RFE).X.glm is less accurate than Low.Cor.X.glm
# check NA coefficients & filter appropriate terms in indep_vars_vctr
# if (method == "glm") {
# orig_glm <- glb_models_lst[[paste0(mdl_id, ".", model_method)]]$finalModel
# orig_glm <- glb_models_lst[["All.X.glm"]]$finalModel; print(summary(orig_glm))
# orig_glm <- glb_models_lst[["RFE.X.glm"]]$finalModel; print(summary(orig_glm))
# require(car)
# vif_orig_glm <- vif(orig_glm); print(vif_orig_glm)
# # if vif errors out with "there are aliased coefficients in the model"
# alias_orig_glm <- alias(orig_glm); alias_complete_orig_glm <- (alias_orig_glm$Complete > 0); alias_complete_orig_glm <- alias_complete_orig_glm[rowSums(alias_complete_orig_glm) > 0, colSums(alias_complete_orig_glm) > 0]; print(alias_complete_orig_glm)
# print(vif_orig_glm[!is.na(vif_orig_glm) & (vif_orig_glm == Inf)])
# print(which.max(vif_orig_glm))
# print(sort(vif_orig_glm[vif_orig_glm >= 1.0e+03], decreasing=TRUE))
# glb_fitobs_df[c(1143, 3637, 3953, 4105), c("UniqueID", "Popular", "H.P.quandary", "Headline")]
# glb_feats_df[glb_feats_df$id %in% grep("[HSA]\\.chrs.n.log", glb_feats_df$id, value=TRUE) | glb_feats_df$cor.high.X %in% grep("[HSA]\\.chrs.n.log", glb_feats_df$id, value=TRUE), ]
# all.equal(glb_allobs_df$S.chrs.uppr.n.log, glb_allobs_df$A.chrs.uppr.n.log)
# cor(glb_allobs_df$S.T.herald, glb_allobs_df$S.T.tribun)
# mydsp_obs(Abstract.contains="[Dd]iar", cols=("Abstract"), all=TRUE)
# subset(glb_feats_df, cor.y.abs <= glb_feats_df[glb_feats_df$id == ".rnorm", "cor.y.abs"])
# corxx_mtrx <- cor(data.matrix(glb_allobs_df[, setdiff(names(glb_allobs_df), myfind_chr_cols_df(glb_allobs_df))]), use="pairwise.complete.obs"); abs_corxx_mtrx <- abs(corxx_mtrx); diag(abs_corxx_mtrx) <- 0
# which.max(abs_corxx_mtrx["S.T.tribun", ])
# abs_corxx_mtrx["A.npnct08.log", "S.npnct08.log"]
# step_glm <- step(orig_glm)
# }
# Since caret does not optimize rpart well
# if (method == "rpart")
# ret_lst <- myfit_mdl(mdl_id=paste0(mdl_id_pfx, ".cp.0"), model_method=method,
# indep_vars_vctr=indep_vars_vctr,
# model_type=glb_model_type,
# rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
# fit_df=glb_fitobs_df, OOB_df=glb_OOBobs_df,
# n_cv_folds=0, tune_models_df=data.frame(parameter="cp", min=0.0, max=0.0, by=0.1))
# User specified
# Ensure at least 2 vars in each regression; else varImp crashes
# sav_models_lst <- glb_models_lst; sav_models_df <- glb_models_df; sav_featsimp_df <- glb_featsimp_df; all.equal(sav_featsimp_df, glb_featsimp_df)
# glb_models_lst <- sav_models_lst; glb_models_df <- sav_models_df; glm_featsimp_df <- sav_featsimp_df
# easier to exclude features
# require(gdata) # needed for trim
# mdl_id <- "";
# indep_vars_vctr <- head(subset(glb_models_df, grepl("All\\.X\\.", mdl_id), select=feats)
# , 1)[, "feats"]
# indep_vars_vctr <- trim(unlist(strsplit(indep_vars_vctr, "[,]")))
# indep_vars_vctr <- setdiff(indep_vars_vctr, ".rnorm")
# easier to include features
#stop(here"); sav_models_df <- glb_models_df; glb_models_df <- sav_models_df
# !_sp
# mdl_id <- "csm"; indep_vars_vctr <- c(NULL
# ,"prdline.my.fctr", "prdline.my.fctr:.clusterid.fctr"
# ,"prdline.my.fctr*biddable"
# #,"prdline.my.fctr*startprice.log"
# #,"prdline.my.fctr*startprice.diff"
# ,"prdline.my.fctr*condition.fctr"
# ,"prdline.my.fctr*D.terms.post.stop.n"
# #,"prdline.my.fctr*D.terms.post.stem.n"
# ,"prdline.my.fctr*cellular.fctr"
# # ,"<feat1>:<feat2>"
# )
# for (method in glbMdlMethods) {
# ret_lst <- myfit_mdl(mdl_id=mdl_id, model_method=method,
# indep_vars_vctr=indep_vars_vctr,
# model_type=glb_model_type,
# rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
# fit_df=glb_fitobs_df, OOB_df=glb_OOBobs_df,
# n_cv_folds=glb_rcv_n_folds, tune_models_df=glb_tune_models_df)
# csm_mdl_id <- paste0(mdl_id, ".", method)
# csm_featsimp_df <- myget_feats_importance(glb_models_lst[[paste0(mdl_id, ".",
# method)]]); print(head(csm_featsimp_df))
# }
###
# Ntv.1.lm <- lm(reformulate(indep_vars_vctr, glb_rsp_var), glb_trnobs_df); print(summary(Ntv.1.lm))
#glb_models_df[, "max.Accuracy.OOB", FALSE]
#varImp(glb_models_lst[["Low.cor.X.glm"]])
#orderBy(~ -Overall, varImp(glb_models_lst[["All.X.2.glm"]])$importance)
#orderBy(~ -Overall, varImp(glb_models_lst[["All.X.3.glm"]])$importance)
#glb_feats_df[grepl("npnct28", glb_feats_df$id), ]
# User specified bivariate models
# indep_vars_vctr_lst <- list()
# for (feat in setdiff(names(glb_fitobs_df),
# union(glb_rsp_var, glbFeatsExclude)))
# indep_vars_vctr_lst[["feat"]] <- feat
# User specified combinatorial models
# indep_vars_vctr_lst <- list()
# combn_mtrx <- combn(c("<feat1_name>", "<feat2_name>", "<featn_name>"),
# <num_feats_to_choose>)
# for (combn_ix in 1:ncol(combn_mtrx))
# #print(combn_mtrx[, combn_ix])
# indep_vars_vctr_lst[[combn_ix]] <- combn_mtrx[, combn_ix]
# template for myfit_mdl
# rf is hard-coded in caret to recognize only Accuracy / Kappa evaluation metrics
# only for OOB in trainControl ?
# ret_lst <- myfit_mdl_fn(mdl_id=paste0(mdl_id_pfx, ""), model_method=method,
# indep_vars_vctr=indep_vars_vctr,
# rsp_var=glb_rsp_var, rsp_var_out=glb_rsp_var_out,
# fit_df=glb_fitobs_df, OOB_df=glb_OOBobs_df,
# n_cv_folds=glb_rcv_n_folds, tune_models_df=glb_tune_models_df,
# model_loss_mtrx=glbMdlMetric_terms,
# model_summaryFunction=glbMdlMetricSummaryFn,
# model_metric=glbMdlMetricSummary,
# model_metric_maximize=glbMdlMetricMaximize)
# Simplify a model
# fit_df <- glb_fitobs_df; glb_mdl <- step(<complex>_mdl)
# Non-caret models
# rpart_area_mdl <- rpart(reformulate("Area", response=glb_rsp_var),
# data=glb_fitobs_df, #method="class",
# control=rpart.control(cp=0.12),
# parms=list(loss=glbMdlMetric_terms))
# print("rpart_sel_wlm_mdl"); prp(rpart_sel_wlm_mdl)
#
print(glb_models_df)## id
## MFO.myMFO_classfr MFO.myMFO_classfr
## Random.myrandom_classfr Random.myrandom_classfr
## Max.cor.Y.rcv.1X1.glmnet Max.cor.Y.rcv.1X1.glmnet
## Max.cor.Y.rcv.3X1.glmnet Max.cor.Y.rcv.3X1.glmnet
## Max.cor.Y.rcv.3X3.glmnet Max.cor.Y.rcv.3X3.glmnet
## Max.cor.Y.rcv.3X5.glmnet Max.cor.Y.rcv.3X5.glmnet
## Max.cor.Y.rcv.5X1.glmnet Max.cor.Y.rcv.5X1.glmnet
## Max.cor.Y.rcv.5X3.glmnet Max.cor.Y.rcv.5X3.glmnet
## Max.cor.Y.rcv.5X5.glmnet Max.cor.Y.rcv.5X5.glmnet
## Max.cor.Y.rcv.1X1.cp.0.rpart Max.cor.Y.rcv.1X1.cp.0.rpart
## Max.cor.Y.rpart Max.cor.Y.rpart
## Interact.High.cor.Y.glmnet Interact.High.cor.Y.glmnet
## Low.cor.X.glmnet Low.cor.X.glmnet
## RFE.X.glmnet RFE.X.glmnet
## RFE.X.glm RFE.X.glm
## All.X.glmnet All.X.glmnet
## Max.cor.Y.rcv.1X1.Interact.glmnet Max.cor.Y.rcv.1X1.Interact.glmnet
## feats
## MFO.myMFO_classfr .rnorm
## Random.myrandom_classfr .rnorm
## Max.cor.Y.rcv.1X1.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.3X1.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.3X3.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.3X5.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.5X1.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.5X3.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.5X5.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.1X1.cp.0.rpart biddable,sprice.root2
## Max.cor.Y.rpart biddable,sprice.root2
## Interact.High.cor.Y.glmnet biddable,sprice.root2,biddable:sprice.log10,biddable:D.terms.post.stop.n.log,biddable:startprice.dcm2.is9,biddable:D.weight.post.stem.sum,biddable:D.chrs.n.log,biddable:cellular.fctr,biddable:D.terms.post.stem.n.log,biddable:sprice.root2
## Low.cor.X.glmnet biddable,spdiff.cut.fctr,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,storage.fctr,D.chrs.pnct11.n.log,startprice.dgt2.is9,color.fctr,prdl.my.fctr,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.root2,prdl.my.fctr:.clusterid.fctr
## RFE.X.glmnet sprice.d20nexp,sprice.log10,sprice.root2,biddable,spdiff.cut.fctr,cellular.fctr,prdl.my.fctr,D.ratio.wrds.stop.n.wrds.n,storage.fctr,condition.fctr,D.weight.sum.stem.stop.Ratio,color.fctr,startprice.dcm2.is9,.rnorm,D.chrs.pnct11.n.log,D.wrds.stop.n.log,startprice.dcm1.is9,startprice.dgt2.is9,D.ratio.weight.sum.wrds.n,D.chrs.pnct13.n.log,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,D.wrds.n.log,D.chrs.uppr.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,D.terms.post.stem.n.log,D.chrs.n.log,startprice.dgt1.is9,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr
## RFE.X.glm sprice.d20nexp,sprice.log10,sprice.root2,biddable,spdiff.cut.fctr,cellular.fctr,prdl.my.fctr,D.ratio.wrds.stop.n.wrds.n,storage.fctr,condition.fctr,D.weight.sum.stem.stop.Ratio,color.fctr,startprice.dcm2.is9,.rnorm,D.chrs.pnct11.n.log,D.wrds.stop.n.log,startprice.dcm1.is9,startprice.dgt2.is9,D.ratio.weight.sum.wrds.n,D.chrs.pnct13.n.log,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,D.wrds.n.log,D.chrs.uppr.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,D.terms.post.stem.n.log,D.chrs.n.log,startprice.dgt1.is9,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr
## All.X.glmnet biddable,sprice.d20nexp,spdiff.cut.fctr,D.ratio.wrds.stop.n.wrds.n,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,startprice.dcm1.is9,storage.fctr,D.chrs.pnct11.n.log,D.chrs.pnct13.n.log,startprice.dgt2.is9,color.fctr,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,prdl.my.fctr,D.chrs.uppr.n.log,D.chrs.n.log,D.wrds.n.log,D.terms.post.stem.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.log10,sprice.root2,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr
## Max.cor.Y.rcv.1X1.Interact.glmnet sprice.root2,biddable
## max.nTuningRuns
## MFO.myMFO_classfr 0
## Random.myrandom_classfr 0
## Max.cor.Y.rcv.1X1.glmnet 0
## Max.cor.Y.rcv.3X1.glmnet 25
## Max.cor.Y.rcv.3X3.glmnet 25
## Max.cor.Y.rcv.3X5.glmnet 25
## Max.cor.Y.rcv.5X1.glmnet 25
## Max.cor.Y.rcv.5X3.glmnet 25
## Max.cor.Y.rcv.5X5.glmnet 25
## Max.cor.Y.rcv.1X1.cp.0.rpart 0
## Max.cor.Y.rpart 5
## Interact.High.cor.Y.glmnet 25
## Low.cor.X.glmnet 25
## RFE.X.glmnet 25
## RFE.X.glm 1
## All.X.glmnet 25
## Max.cor.Y.rcv.1X1.Interact.glmnet 25
## min.elapsedtime.everything
## MFO.myMFO_classfr 0.292
## Random.myrandom_classfr 0.308
## Max.cor.Y.rcv.1X1.glmnet 0.740
## Max.cor.Y.rcv.3X1.glmnet 1.546
## Max.cor.Y.rcv.3X3.glmnet 2.078
## Max.cor.Y.rcv.3X5.glmnet 2.568
## Max.cor.Y.rcv.5X1.glmnet 1.729
## Max.cor.Y.rcv.5X3.glmnet 2.597
## Max.cor.Y.rcv.5X5.glmnet 3.111
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.750
## Max.cor.Y.rpart 1.632
## Interact.High.cor.Y.glmnet 2.747
## Low.cor.X.glmnet 3.129
## RFE.X.glmnet 15.219
## RFE.X.glm 2.890
## All.X.glmnet 14.826
## Max.cor.Y.rcv.1X1.Interact.glmnet 2.319
## min.elapsedtime.final max.AUCpROC.fit
## MFO.myMFO_classfr 0.003 0.5000000
## Random.myrandom_classfr 0.001 0.4859659
## Max.cor.Y.rcv.1X1.glmnet 0.023 0.7898522
## Max.cor.Y.rcv.3X1.glmnet 0.018 0.7919708
## Max.cor.Y.rcv.3X3.glmnet 0.016 0.7919708
## Max.cor.Y.rcv.3X5.glmnet 0.016 0.7919708
## Max.cor.Y.rcv.5X1.glmnet 0.015 0.7956204
## Max.cor.Y.rcv.5X3.glmnet 0.016 0.7956204
## Max.cor.Y.rcv.5X5.glmnet 0.015 0.7963859
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.012 0.8281424
## Max.cor.Y.rpart 0.013 0.8102808
## Interact.High.cor.Y.glmnet 0.100 0.7970973
## Low.cor.X.glmnet 0.121 0.8597751
## RFE.X.glmnet 0.534 0.8626593
## RFE.X.glm 0.192 0.8597751
## All.X.glmnet 0.545 0.8626593
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.016 0.7919708
## max.Sens.fit max.Spec.fit
## MFO.myMFO_classfr 1.0000000 0.0000000
## Random.myrandom_classfr 0.5164234 0.4555085
## Max.cor.Y.rcv.1X1.glmnet 0.8339416 0.7457627
## Max.cor.Y.rcv.3X1.glmnet 0.8339416 0.7500000
## Max.cor.Y.rcv.3X3.glmnet 0.8339416 0.7500000
## Max.cor.Y.rcv.3X5.glmnet 0.8339416 0.7500000
## Max.cor.Y.rcv.5X1.glmnet 0.8412409 0.7500000
## Max.cor.Y.rcv.5X3.glmnet 0.8412409 0.7500000
## Max.cor.Y.rcv.5X5.glmnet 0.8448905 0.7478814
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.8978102 0.7584746
## Max.cor.Y.rpart 0.8959854 0.7245763
## Interact.High.cor.Y.glmnet 0.8886861 0.7055085
## Low.cor.X.glmnet 0.8996350 0.8199153
## RFE.X.glmnet 0.9032847 0.8220339
## RFE.X.glm 0.8996350 0.8199153
## All.X.glmnet 0.9032847 0.8220339
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.8339416 0.7500000
## max.AUCROCR.fit opt.prob.threshold.fit
## MFO.myMFO_classfr 0.5000000 0.4
## Random.myrandom_classfr 0.4893217 0.4
## Max.cor.Y.rcv.1X1.glmnet 0.8620600 0.4
## Max.cor.Y.rcv.3X1.glmnet 0.8621335 0.5
## Max.cor.Y.rcv.3X3.glmnet 0.8622147 0.5
## Max.cor.Y.rcv.3X5.glmnet 0.8622069 0.5
## Max.cor.Y.rcv.5X1.glmnet 0.8621064 0.5
## Max.cor.Y.rcv.5X3.glmnet 0.8621064 0.5
## Max.cor.Y.rcv.5X5.glmnet 0.8621064 0.5
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.8816304 0.4
## Max.cor.Y.rpart 0.8387008 0.5
## Interact.High.cor.Y.glmnet 0.8639370 0.4
## Low.cor.X.glmnet 0.9329264 0.5
## RFE.X.glmnet 0.9355785 0.4
## RFE.X.glm 0.9392823 0.4
## All.X.glmnet 0.9355785 0.4
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.8622147 0.5
## max.f.score.fit max.Accuracy.fit
## MFO.myMFO_classfr 0.6327078 0.4627451
## Random.myrandom_classfr 0.6327078 0.4627451
## Max.cor.Y.rcv.1X1.glmnet 0.7698745 0.7843137
## Max.cor.Y.rcv.3X1.glmnet 0.7720829 0.7950911
## Max.cor.Y.rcv.3X3.glmnet 0.7720829 0.7954221
## Max.cor.Y.rcv.3X5.glmnet 0.7720829 0.7962655
## Max.cor.Y.rcv.5X1.glmnet 0.7754655 0.7990899
## Max.cor.Y.rcv.5X3.glmnet 0.7754655 0.7964364
## Max.cor.Y.rcv.5X5.glmnet 0.7758242 0.7970687
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.8102109 0.8323529
## Max.cor.Y.rpart 0.7853042 0.7964073
## Interact.High.cor.Y.glmnet 0.7724750 0.8026098
## Low.cor.X.glmnet 0.8468271 0.8349657
## RFE.X.glmnet 0.8503119 0.8303886
## RFE.X.glm 0.8496868 0.8307135
## All.X.glmnet 0.8503119 0.8303886
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.7720829 0.7954221
## max.AccuracyLower.fit
## MFO.myMFO_classfr 0.4318011
## Random.myrandom_classfr 0.4318011
## Max.cor.Y.rcv.1X1.glmnet 0.7577802
## Max.cor.Y.rcv.3X1.glmnet 0.7690020
## Max.cor.Y.rcv.3X3.glmnet 0.7690020
## Max.cor.Y.rcv.3X5.glmnet 0.7690020
## Max.cor.Y.rcv.5X1.glmnet 0.7730893
## Max.cor.Y.rcv.5X3.glmnet 0.7730893
## Max.cor.Y.rcv.5X5.glmnet 0.7741117
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.8079872
## Max.cor.Y.rpart 0.7915283
## Interact.High.cor.Y.glmnet 0.7730893
## Low.cor.X.glmnet 0.8400893
## RFE.X.glmnet 0.8359295
## RFE.X.glm 0.8359295
## All.X.glmnet 0.8359295
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.7690020
## max.AccuracyUpper.fit max.Kappa.fit
## MFO.myMFO_classfr 0.4939047 0.0000000
## Random.myrandom_classfr 0.4939047 0.0000000
## Max.cor.Y.rcv.1X1.glmnet 0.8091950 0.5669826
## Max.cor.Y.rcv.3X1.glmnet 0.8194783 0.5860164
## Max.cor.Y.rcv.3X3.glmnet 0.8194783 0.5866808
## Max.cor.Y.rcv.3X5.glmnet 0.8194783 0.5883400
## Max.cor.Y.rcv.5X1.glmnet 0.8232111 0.5944245
## Max.cor.Y.rcv.5X3.glmnet 0.8232111 0.5888026
## Max.cor.Y.rcv.5X5.glmnet 0.8241437 0.5901123
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.8547824 0.6606905
## Max.cor.Y.rpart 0.8399619 0.5852116
## Interact.High.cor.Y.glmnet 0.8232111 0.5988972
## Low.cor.X.glmnet 0.8832827 0.6666179
## RFE.X.glmnet 0.8796230 0.6576319
## RFE.X.glm 0.8796230 0.6585198
## All.X.glmnet 0.8796230 0.6576319
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.8194783 0.5866808
## max.AUCpROC.OOB max.Sens.OOB
## MFO.myMFO_classfr 0.5000000 1.0000000
## Random.myrandom_classfr 0.5064252 0.5480984
## Max.cor.Y.rcv.1X1.glmnet 0.7665376 0.8255034
## Max.cor.Y.rcv.3X1.glmnet 0.7592654 0.8187919
## Max.cor.Y.rcv.3X3.glmnet 0.7603840 0.8210291
## Max.cor.Y.rcv.3X5.glmnet 0.7603840 0.8210291
## Max.cor.Y.rcv.5X1.glmnet 0.7615026 0.8232662
## Max.cor.Y.rcv.5X3.glmnet 0.7615026 0.8232662
## Max.cor.Y.rcv.5X5.glmnet 0.7641135 0.8232662
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.7504950 0.8456376
## Max.cor.Y.rpart 0.7445254 0.8545861
## Interact.High.cor.Y.glmnet 0.7551533 0.8680089
## Low.cor.X.glmnet 0.7801532 0.8187919
## RFE.X.glmnet 0.7835089 0.8255034
## RFE.X.glm 0.7851882 0.8210291
## All.X.glmnet 0.7835089 0.8255034
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.7603840 0.8210291
## max.Spec.OOB max.AUCROCR.OOB
## MFO.myMFO_classfr 0.0000000 0.5000000
## Random.myrandom_classfr 0.4647520 0.4956046
## Max.cor.Y.rcv.1X1.glmnet 0.7075718 0.8197236
## Max.cor.Y.rcv.3X1.glmnet 0.6997389 0.8197353
## Max.cor.Y.rcv.3X3.glmnet 0.6997389 0.8197470
## Max.cor.Y.rcv.3X5.glmnet 0.6997389 0.8196593
## Max.cor.Y.rcv.5X1.glmnet 0.6997389 0.8195600
## Max.cor.Y.rcv.5X3.glmnet 0.6997389 0.8195600
## Max.cor.Y.rcv.5X5.glmnet 0.7049608 0.8193790
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.6553525 0.7934621
## Max.cor.Y.rpart 0.6344648 0.7855854
## Interact.High.cor.Y.glmnet 0.6422977 0.8189000
## Low.cor.X.glmnet 0.7415144 0.8727052
## RFE.X.glmnet 0.7415144 0.8721444
## RFE.X.glm 0.7493473 0.8700475
## All.X.glmnet 0.7415144 0.8721444
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.6997389 0.8197470
## opt.prob.threshold.OOB max.f.score.OOB
## MFO.myMFO_classfr 0.4 0.6314922
## Random.myrandom_classfr 0.4 0.6314922
## Max.cor.Y.rcv.1X1.glmnet 0.4 0.7535854
## Max.cor.Y.rcv.3X1.glmnet 0.4 0.7519182
## Max.cor.Y.rcv.3X3.glmnet 0.4 0.7525510
## Max.cor.Y.rcv.3X5.glmnet 0.4 0.7515924
## Max.cor.Y.rcv.5X1.glmnet 0.4 0.7506361
## Max.cor.Y.rcv.5X3.glmnet 0.4 0.7506361
## Max.cor.Y.rcv.5X5.glmnet 0.4 0.7512690
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.3 0.7313997
## Max.cor.Y.rpart 0.3 0.7394270
## Interact.High.cor.Y.glmnet 0.3 0.7522698
## Low.cor.X.glmnet 0.3 0.7719715
## RFE.X.glmnet 0.3 0.7654028
## RFE.X.glm 0.5 0.7653333
## All.X.glmnet 0.3 0.7654028
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.4 0.7525510
## max.Accuracy.OOB max.AccuracyLower.OOB
## MFO.myMFO_classfr 0.4614458 0.4271177
## Random.myrandom_classfr 0.4614458 0.4271177
## Max.cor.Y.rcv.1X1.glmnet 0.7722892 0.7422177
## Max.cor.Y.rcv.3X1.glmnet 0.7662651 0.7359543
## Max.cor.Y.rcv.3X3.glmnet 0.7662651 0.7359543
## Max.cor.Y.rcv.3X5.glmnet 0.7650602 0.7347026
## Max.cor.Y.rcv.5X1.glmnet 0.7638554 0.7334512
## Max.cor.Y.rcv.5X3.glmnet 0.7638554 0.7334512
## Max.cor.Y.rcv.5X5.glmnet 0.7638554 0.7334512
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.7433735 0.7122254
## Max.cor.Y.rpart 0.7698795 0.7397113
## Interact.High.cor.Y.glmnet 0.7698795 0.7397113
## Low.cor.X.glmnet 0.7686747 0.7384586
## RFE.X.glmnet 0.7614458 0.7309494
## RFE.X.glm 0.7879518 0.7585420
## All.X.glmnet 0.7614458 0.7309494
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.7662651 0.7359543
## max.AccuracyUpper.OOB max.Kappa.OOB
## MFO.myMFO_classfr 0.4960486 0.0000000
## Random.myrandom_classfr 0.4960486 0.0000000
## Max.cor.Y.rcv.1X1.glmnet 0.8004125 0.5419399
## Max.cor.Y.rcv.3X1.glmnet 0.7946712 0.5311363
## Max.cor.Y.rcv.3X3.glmnet 0.7946712 0.5313109
## Max.cor.Y.rcv.3X5.glmnet 0.7935220 0.5289828
## Max.cor.Y.rcv.5X1.glmnet 0.7923725 0.5266555
## Max.cor.Y.rcv.5X3.glmnet 0.7923725 0.5266555
## Max.cor.Y.rcv.5X5.glmnet 0.7923725 0.5268317
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.7727821 0.4862697
## Max.cor.Y.rpart 0.7981170 0.5341327
## Interact.High.cor.Y.glmnet 0.7981170 0.5374385
## Low.cor.X.glmnet 0.7969687 0.5411011
## RFE.X.glmnet 0.7900724 0.5269348
## RFE.X.glm 0.8152999 0.5720872
## All.X.glmnet 0.7900724 0.5269348
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.7946712 0.5313109
## max.AccuracySD.fit max.KappaSD.fit
## MFO.myMFO_classfr NA NA
## Random.myrandom_classfr NA NA
## Max.cor.Y.rcv.1X1.glmnet NA NA
## Max.cor.Y.rcv.3X1.glmnet 0.003934893 0.009324777
## Max.cor.Y.rcv.3X3.glmnet 0.020088705 0.040878031
## Max.cor.Y.rcv.3X5.glmnet 0.018120665 0.036897112
## Max.cor.Y.rcv.5X1.glmnet 0.036086436 0.070687264
## Max.cor.Y.rcv.5X3.glmnet 0.033491865 0.067292776
## Max.cor.Y.rcv.5X5.glmnet 0.031290870 0.062857472
## Max.cor.Y.rcv.1X1.cp.0.rpart NA NA
## Max.cor.Y.rpart 0.020068991 0.040059154
## Interact.High.cor.Y.glmnet 0.017969767 0.037056533
## Low.cor.X.glmnet 0.009378705 0.019384018
## RFE.X.glmnet 0.011266493 0.022766184
## RFE.X.glm 0.014477821 0.029001944
## All.X.glmnet 0.011266493 0.022766184
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.020088705 0.040878031rm(ret_lst)
fit.models_1_chunk_df <- myadd_chunk(fit.models_1_chunk_df, "fit.models_1_end",
major.inc=TRUE, label.minor="teardown")## label step_major step_minor
## 8 fit.models_1_Max.cor.Y.rcv.1X1.Interact 4 1
## 9 fit.models_1_end 5 0
## label_minor bgn end elapsed
## 8 glmnet 470.840 477.204 6.365
## 9 teardown 477.205 NA NAglb_chunks_df <- myadd_chunk(glb_chunks_df, "fit.models", major.inc=FALSE)## label step_major step_minor label_minor bgn end elapsed
## 11 fit.models 6 1 1 409.619 477.214 67.595
## 12 fit.models 6 2 2 477.215 NA NA#stop(here"); glb_to_sav(); all.equal(glb_models_df, sav_models_df)
# if (!is.null(glbMdlMetricSummaryFn)) {
# stats_df <- glb_models_df[, "id", FALSE]
#
# stats_mdl_df <- data.frame()
# for (mdl_id in stats_df$id) {
# stats_mdl_df <- rbind(stats_mdl_df,
# mypredict_mdl(glb_models_lst[[mdl_id]], glb_fitobs_df, glb_rsp_var,
# glb_rsp_var_out, mdl_id, "fit",
# glbMdlMetricSummaryFn, glbMdlMetricSummary,
# glbMdlMetricMaximize, ret_type="stats"))
# }
# stats_df <- merge(stats_df, stats_mdl_df, all.x=TRUE)
#
# stats_mdl_df <- data.frame()
# for (mdl_id in stats_df$id) {
# stats_mdl_df <- rbind(stats_mdl_df,
# mypredict_mdl(glb_models_lst[[mdl_id]], glb_OOBobs_df, glb_rsp_var,
# glb_rsp_var_out, mdl_id, "OOB",
# glbMdlMetricSummaryFn, glbMdlMetricSummary,
# glbMdlMetricMaximize, ret_type="stats"))
# }
# stats_df <- merge(stats_df, stats_mdl_df, all.x=TRUE)
#
# print("Merging following data into glb_models_df:")
# print(stats_mrg_df <- stats_df[, c(1, grep(glbMdlMetricSummary, names(stats_df)))])
# print(tmp_models_df <- orderBy(~id, glb_models_df[, c("id",
# grep(glbMdlMetricSummary, names(stats_df), value=TRUE))]))
#
# tmp2_models_df <- glb_models_df[, c("id", setdiff(names(glb_models_df),
# grep(glbMdlMetricSummary, names(stats_df), value=TRUE)))]
# tmp3_models_df <- merge(tmp2_models_df, stats_mrg_df, all.x=TRUE, sort=FALSE)
# print(tmp3_models_df)
# print(names(tmp3_models_df))
# print(glb_models_df <- subset(tmp3_models_df, select=-id.1))
# }
plt_models_df <- glb_models_df[, -grep("SD|Upper|Lower", names(glb_models_df))]
for (var in grep("^min.", names(plt_models_df), value=TRUE)) {
plt_models_df[, sub("min.", "inv.", var)] <-
#ifelse(all(is.na(tmp <- plt_models_df[, var])), NA, 1.0 / tmp)
1.0 / plt_models_df[, var]
plt_models_df <- plt_models_df[ , -grep(var, names(plt_models_df))]
}
print(plt_models_df)## id
## MFO.myMFO_classfr MFO.myMFO_classfr
## Random.myrandom_classfr Random.myrandom_classfr
## Max.cor.Y.rcv.1X1.glmnet Max.cor.Y.rcv.1X1.glmnet
## Max.cor.Y.rcv.3X1.glmnet Max.cor.Y.rcv.3X1.glmnet
## Max.cor.Y.rcv.3X3.glmnet Max.cor.Y.rcv.3X3.glmnet
## Max.cor.Y.rcv.3X5.glmnet Max.cor.Y.rcv.3X5.glmnet
## Max.cor.Y.rcv.5X1.glmnet Max.cor.Y.rcv.5X1.glmnet
## Max.cor.Y.rcv.5X3.glmnet Max.cor.Y.rcv.5X3.glmnet
## Max.cor.Y.rcv.5X5.glmnet Max.cor.Y.rcv.5X5.glmnet
## Max.cor.Y.rcv.1X1.cp.0.rpart Max.cor.Y.rcv.1X1.cp.0.rpart
## Max.cor.Y.rpart Max.cor.Y.rpart
## Interact.High.cor.Y.glmnet Interact.High.cor.Y.glmnet
## Low.cor.X.glmnet Low.cor.X.glmnet
## RFE.X.glmnet RFE.X.glmnet
## RFE.X.glm RFE.X.glm
## All.X.glmnet All.X.glmnet
## Max.cor.Y.rcv.1X1.Interact.glmnet Max.cor.Y.rcv.1X1.Interact.glmnet
## feats
## MFO.myMFO_classfr .rnorm
## Random.myrandom_classfr .rnorm
## Max.cor.Y.rcv.1X1.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.3X1.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.3X3.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.3X5.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.5X1.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.5X3.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.5X5.glmnet biddable,sprice.root2
## Max.cor.Y.rcv.1X1.cp.0.rpart biddable,sprice.root2
## Max.cor.Y.rpart biddable,sprice.root2
## Interact.High.cor.Y.glmnet biddable,sprice.root2,biddable:sprice.log10,biddable:D.terms.post.stop.n.log,biddable:startprice.dcm2.is9,biddable:D.weight.post.stem.sum,biddable:D.chrs.n.log,biddable:cellular.fctr,biddable:D.terms.post.stem.n.log,biddable:sprice.root2
## Low.cor.X.glmnet biddable,spdiff.cut.fctr,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,storage.fctr,D.chrs.pnct11.n.log,startprice.dgt2.is9,color.fctr,prdl.my.fctr,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.root2,prdl.my.fctr:.clusterid.fctr
## RFE.X.glmnet sprice.d20nexp,sprice.log10,sprice.root2,biddable,spdiff.cut.fctr,cellular.fctr,prdl.my.fctr,D.ratio.wrds.stop.n.wrds.n,storage.fctr,condition.fctr,D.weight.sum.stem.stop.Ratio,color.fctr,startprice.dcm2.is9,.rnorm,D.chrs.pnct11.n.log,D.wrds.stop.n.log,startprice.dcm1.is9,startprice.dgt2.is9,D.ratio.weight.sum.wrds.n,D.chrs.pnct13.n.log,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,D.wrds.n.log,D.chrs.uppr.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,D.terms.post.stem.n.log,D.chrs.n.log,startprice.dgt1.is9,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr
## RFE.X.glm sprice.d20nexp,sprice.log10,sprice.root2,biddable,spdiff.cut.fctr,cellular.fctr,prdl.my.fctr,D.ratio.wrds.stop.n.wrds.n,storage.fctr,condition.fctr,D.weight.sum.stem.stop.Ratio,color.fctr,startprice.dcm2.is9,.rnorm,D.chrs.pnct11.n.log,D.wrds.stop.n.log,startprice.dcm1.is9,startprice.dgt2.is9,D.ratio.weight.sum.wrds.n,D.chrs.pnct13.n.log,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,D.wrds.n.log,D.chrs.uppr.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,D.terms.post.stem.n.log,D.chrs.n.log,startprice.dgt1.is9,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr
## All.X.glmnet biddable,sprice.d20nexp,spdiff.cut.fctr,D.ratio.wrds.stop.n.wrds.n,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,startprice.dcm1.is9,storage.fctr,D.chrs.pnct11.n.log,D.chrs.pnct13.n.log,startprice.dgt2.is9,color.fctr,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,prdl.my.fctr,D.chrs.uppr.n.log,D.chrs.n.log,D.wrds.n.log,D.terms.post.stem.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.log10,sprice.root2,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr
## Max.cor.Y.rcv.1X1.Interact.glmnet sprice.root2,biddable
## max.nTuningRuns max.AUCpROC.fit
## MFO.myMFO_classfr 0 0.5000000
## Random.myrandom_classfr 0 0.4859659
## Max.cor.Y.rcv.1X1.glmnet 0 0.7898522
## Max.cor.Y.rcv.3X1.glmnet 25 0.7919708
## Max.cor.Y.rcv.3X3.glmnet 25 0.7919708
## Max.cor.Y.rcv.3X5.glmnet 25 0.7919708
## Max.cor.Y.rcv.5X1.glmnet 25 0.7956204
## Max.cor.Y.rcv.5X3.glmnet 25 0.7956204
## Max.cor.Y.rcv.5X5.glmnet 25 0.7963859
## Max.cor.Y.rcv.1X1.cp.0.rpart 0 0.8281424
## Max.cor.Y.rpart 5 0.8102808
## Interact.High.cor.Y.glmnet 25 0.7970973
## Low.cor.X.glmnet 25 0.8597751
## RFE.X.glmnet 25 0.8626593
## RFE.X.glm 1 0.8597751
## All.X.glmnet 25 0.8626593
## Max.cor.Y.rcv.1X1.Interact.glmnet 25 0.7919708
## max.Sens.fit max.Spec.fit
## MFO.myMFO_classfr 1.0000000 0.0000000
## Random.myrandom_classfr 0.5164234 0.4555085
## Max.cor.Y.rcv.1X1.glmnet 0.8339416 0.7457627
## Max.cor.Y.rcv.3X1.glmnet 0.8339416 0.7500000
## Max.cor.Y.rcv.3X3.glmnet 0.8339416 0.7500000
## Max.cor.Y.rcv.3X5.glmnet 0.8339416 0.7500000
## Max.cor.Y.rcv.5X1.glmnet 0.8412409 0.7500000
## Max.cor.Y.rcv.5X3.glmnet 0.8412409 0.7500000
## Max.cor.Y.rcv.5X5.glmnet 0.8448905 0.7478814
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.8978102 0.7584746
## Max.cor.Y.rpart 0.8959854 0.7245763
## Interact.High.cor.Y.glmnet 0.8886861 0.7055085
## Low.cor.X.glmnet 0.8996350 0.8199153
## RFE.X.glmnet 0.9032847 0.8220339
## RFE.X.glm 0.8996350 0.8199153
## All.X.glmnet 0.9032847 0.8220339
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.8339416 0.7500000
## max.AUCROCR.fit opt.prob.threshold.fit
## MFO.myMFO_classfr 0.5000000 0.4
## Random.myrandom_classfr 0.4893217 0.4
## Max.cor.Y.rcv.1X1.glmnet 0.8620600 0.4
## Max.cor.Y.rcv.3X1.glmnet 0.8621335 0.5
## Max.cor.Y.rcv.3X3.glmnet 0.8622147 0.5
## Max.cor.Y.rcv.3X5.glmnet 0.8622069 0.5
## Max.cor.Y.rcv.5X1.glmnet 0.8621064 0.5
## Max.cor.Y.rcv.5X3.glmnet 0.8621064 0.5
## Max.cor.Y.rcv.5X5.glmnet 0.8621064 0.5
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.8816304 0.4
## Max.cor.Y.rpart 0.8387008 0.5
## Interact.High.cor.Y.glmnet 0.8639370 0.4
## Low.cor.X.glmnet 0.9329264 0.5
## RFE.X.glmnet 0.9355785 0.4
## RFE.X.glm 0.9392823 0.4
## All.X.glmnet 0.9355785 0.4
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.8622147 0.5
## max.f.score.fit max.Accuracy.fit
## MFO.myMFO_classfr 0.6327078 0.4627451
## Random.myrandom_classfr 0.6327078 0.4627451
## Max.cor.Y.rcv.1X1.glmnet 0.7698745 0.7843137
## Max.cor.Y.rcv.3X1.glmnet 0.7720829 0.7950911
## Max.cor.Y.rcv.3X3.glmnet 0.7720829 0.7954221
## Max.cor.Y.rcv.3X5.glmnet 0.7720829 0.7962655
## Max.cor.Y.rcv.5X1.glmnet 0.7754655 0.7990899
## Max.cor.Y.rcv.5X3.glmnet 0.7754655 0.7964364
## Max.cor.Y.rcv.5X5.glmnet 0.7758242 0.7970687
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.8102109 0.8323529
## Max.cor.Y.rpart 0.7853042 0.7964073
## Interact.High.cor.Y.glmnet 0.7724750 0.8026098
## Low.cor.X.glmnet 0.8468271 0.8349657
## RFE.X.glmnet 0.8503119 0.8303886
## RFE.X.glm 0.8496868 0.8307135
## All.X.glmnet 0.8503119 0.8303886
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.7720829 0.7954221
## max.Kappa.fit max.AUCpROC.OOB
## MFO.myMFO_classfr 0.0000000 0.5000000
## Random.myrandom_classfr 0.0000000 0.5064252
## Max.cor.Y.rcv.1X1.glmnet 0.5669826 0.7665376
## Max.cor.Y.rcv.3X1.glmnet 0.5860164 0.7592654
## Max.cor.Y.rcv.3X3.glmnet 0.5866808 0.7603840
## Max.cor.Y.rcv.3X5.glmnet 0.5883400 0.7603840
## Max.cor.Y.rcv.5X1.glmnet 0.5944245 0.7615026
## Max.cor.Y.rcv.5X3.glmnet 0.5888026 0.7615026
## Max.cor.Y.rcv.5X5.glmnet 0.5901123 0.7641135
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.6606905 0.7504950
## Max.cor.Y.rpart 0.5852116 0.7445254
## Interact.High.cor.Y.glmnet 0.5988972 0.7551533
## Low.cor.X.glmnet 0.6666179 0.7801532
## RFE.X.glmnet 0.6576319 0.7835089
## RFE.X.glm 0.6585198 0.7851882
## All.X.glmnet 0.6576319 0.7835089
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.5866808 0.7603840
## max.Sens.OOB max.Spec.OOB
## MFO.myMFO_classfr 1.0000000 0.0000000
## Random.myrandom_classfr 0.5480984 0.4647520
## Max.cor.Y.rcv.1X1.glmnet 0.8255034 0.7075718
## Max.cor.Y.rcv.3X1.glmnet 0.8187919 0.6997389
## Max.cor.Y.rcv.3X3.glmnet 0.8210291 0.6997389
## Max.cor.Y.rcv.3X5.glmnet 0.8210291 0.6997389
## Max.cor.Y.rcv.5X1.glmnet 0.8232662 0.6997389
## Max.cor.Y.rcv.5X3.glmnet 0.8232662 0.6997389
## Max.cor.Y.rcv.5X5.glmnet 0.8232662 0.7049608
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.8456376 0.6553525
## Max.cor.Y.rpart 0.8545861 0.6344648
## Interact.High.cor.Y.glmnet 0.8680089 0.6422977
## Low.cor.X.glmnet 0.8187919 0.7415144
## RFE.X.glmnet 0.8255034 0.7415144
## RFE.X.glm 0.8210291 0.7493473
## All.X.glmnet 0.8255034 0.7415144
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.8210291 0.6997389
## max.AUCROCR.OOB opt.prob.threshold.OOB
## MFO.myMFO_classfr 0.5000000 0.4
## Random.myrandom_classfr 0.4956046 0.4
## Max.cor.Y.rcv.1X1.glmnet 0.8197236 0.4
## Max.cor.Y.rcv.3X1.glmnet 0.8197353 0.4
## Max.cor.Y.rcv.3X3.glmnet 0.8197470 0.4
## Max.cor.Y.rcv.3X5.glmnet 0.8196593 0.4
## Max.cor.Y.rcv.5X1.glmnet 0.8195600 0.4
## Max.cor.Y.rcv.5X3.glmnet 0.8195600 0.4
## Max.cor.Y.rcv.5X5.glmnet 0.8193790 0.4
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.7934621 0.3
## Max.cor.Y.rpart 0.7855854 0.3
## Interact.High.cor.Y.glmnet 0.8189000 0.3
## Low.cor.X.glmnet 0.8727052 0.3
## RFE.X.glmnet 0.8721444 0.3
## RFE.X.glm 0.8700475 0.5
## All.X.glmnet 0.8721444 0.3
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.8197470 0.4
## max.f.score.OOB max.Accuracy.OOB
## MFO.myMFO_classfr 0.6314922 0.4614458
## Random.myrandom_classfr 0.6314922 0.4614458
## Max.cor.Y.rcv.1X1.glmnet 0.7535854 0.7722892
## Max.cor.Y.rcv.3X1.glmnet 0.7519182 0.7662651
## Max.cor.Y.rcv.3X3.glmnet 0.7525510 0.7662651
## Max.cor.Y.rcv.3X5.glmnet 0.7515924 0.7650602
## Max.cor.Y.rcv.5X1.glmnet 0.7506361 0.7638554
## Max.cor.Y.rcv.5X3.glmnet 0.7506361 0.7638554
## Max.cor.Y.rcv.5X5.glmnet 0.7512690 0.7638554
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.7313997 0.7433735
## Max.cor.Y.rpart 0.7394270 0.7698795
## Interact.High.cor.Y.glmnet 0.7522698 0.7698795
## Low.cor.X.glmnet 0.7719715 0.7686747
## RFE.X.glmnet 0.7654028 0.7614458
## RFE.X.glm 0.7653333 0.7879518
## All.X.glmnet 0.7654028 0.7614458
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.7525510 0.7662651
## max.Kappa.OOB inv.elapsedtime.everything
## MFO.myMFO_classfr 0.0000000 3.42465753
## Random.myrandom_classfr 0.0000000 3.24675325
## Max.cor.Y.rcv.1X1.glmnet 0.5419399 1.35135135
## Max.cor.Y.rcv.3X1.glmnet 0.5311363 0.64683053
## Max.cor.Y.rcv.3X3.glmnet 0.5313109 0.48123195
## Max.cor.Y.rcv.3X5.glmnet 0.5289828 0.38940810
## Max.cor.Y.rcv.5X1.glmnet 0.5266555 0.57836900
## Max.cor.Y.rcv.5X3.glmnet 0.5266555 0.38505968
## Max.cor.Y.rcv.5X5.glmnet 0.5268317 0.32144005
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.4862697 1.33333333
## Max.cor.Y.rpart 0.5341327 0.61274510
## Interact.High.cor.Y.glmnet 0.5374385 0.36403349
## Low.cor.X.glmnet 0.5411011 0.31959092
## RFE.X.glmnet 0.5269348 0.06570734
## RFE.X.glm 0.5720872 0.34602076
## All.X.glmnet 0.5269348 0.06744908
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.5313109 0.43122035
## inv.elapsedtime.final
## MFO.myMFO_classfr 333.333333
## Random.myrandom_classfr 1000.000000
## Max.cor.Y.rcv.1X1.glmnet 43.478261
## Max.cor.Y.rcv.3X1.glmnet 55.555556
## Max.cor.Y.rcv.3X3.glmnet 62.500000
## Max.cor.Y.rcv.3X5.glmnet 62.500000
## Max.cor.Y.rcv.5X1.glmnet 66.666667
## Max.cor.Y.rcv.5X3.glmnet 62.500000
## Max.cor.Y.rcv.5X5.glmnet 66.666667
## Max.cor.Y.rcv.1X1.cp.0.rpart 83.333333
## Max.cor.Y.rpart 76.923077
## Interact.High.cor.Y.glmnet 10.000000
## Low.cor.X.glmnet 8.264463
## RFE.X.glmnet 1.872659
## RFE.X.glm 5.208333
## All.X.glmnet 1.834862
## Max.cor.Y.rcv.1X1.Interact.glmnet 62.500000print(myplot_radar(radar_inp_df=plt_models_df))## Warning in RColorBrewer::brewer.pal(n, pal): n too large, allowed maximum for palette Set1 is 9
## Returning the palette you asked for with that many colors## Warning: The shape palette can deal with a maximum of 6 discrete values
## because more than 6 becomes difficult to discriminate; you have
## 17. Consider specifying shapes manually if you must have them.## Warning: Removed 220 rows containing missing values (geom_point).## Warning in RColorBrewer::brewer.pal(n, pal): n too large, allowed maximum for palette Set1 is 9
## Returning the palette you asked for with that many colors## Warning: The shape palette can deal with a maximum of 6 discrete values
## because more than 6 becomes difficult to discriminate; you have
## 17. Consider specifying shapes manually if you must have them.
# print(myplot_radar(radar_inp_df=subset(plt_models_df,
# !(mdl_id %in% grep("random|MFO", plt_models_df$id, value=TRUE)))))
# Compute CI for <metric>SD
glb_models_df <- mutate(glb_models_df,
max.df = ifelse(max.nTuningRuns > 1, max.nTuningRuns - 1, NA),
min.sd2ci.scaler = ifelse(is.na(max.df), NA, qt(0.975, max.df)))
for (var in grep("SD", names(glb_models_df), value=TRUE)) {
# Does CI alredy exist ?
var_components <- unlist(strsplit(var, "SD"))
varActul <- paste0(var_components[1], var_components[2])
varUpper <- paste0(var_components[1], "Upper", var_components[2])
varLower <- paste0(var_components[1], "Lower", var_components[2])
if (varUpper %in% names(glb_models_df)) {
warning(varUpper, " already exists in glb_models_df")
# Assuming Lower also exists
next
}
print(sprintf("var:%s", var))
# CI is dependent on sample size in t distribution; df=n-1
glb_models_df[, varUpper] <- glb_models_df[, varActul] +
glb_models_df[, "min.sd2ci.scaler"] * glb_models_df[, var]
glb_models_df[, varLower] <- glb_models_df[, varActul] -
glb_models_df[, "min.sd2ci.scaler"] * glb_models_df[, var]
}## Warning: max.AccuracyUpper.fit already exists in glb_models_df## [1] "var:max.KappaSD.fit"# Plot metrics with CI
plt_models_df <- glb_models_df[, "id", FALSE]
pltCI_models_df <- glb_models_df[, "id", FALSE]
for (var in grep("Upper", names(glb_models_df), value=TRUE)) {
var_components <- unlist(strsplit(var, "Upper"))
col_name <- unlist(paste(var_components, collapse=""))
plt_models_df[, col_name] <- glb_models_df[, col_name]
for (name in paste0(var_components[1], c("Upper", "Lower"), var_components[2]))
pltCI_models_df[, name] <- glb_models_df[, name]
}
build_statsCI_data <- function(plt_models_df) {
mltd_models_df <- melt(plt_models_df, id.vars="id")
mltd_models_df$data <- sapply(1:nrow(mltd_models_df),
function(row_ix) tail(unlist(strsplit(as.character(
mltd_models_df[row_ix, "variable"]), "[.]")), 1))
mltd_models_df$label <- sapply(1:nrow(mltd_models_df),
function(row_ix) head(unlist(strsplit(as.character(
mltd_models_df[row_ix, "variable"]),
paste0(".", mltd_models_df[row_ix, "data"]))), 1))
#print(mltd_models_df)
return(mltd_models_df)
}
mltd_models_df <- build_statsCI_data(plt_models_df)
mltdCI_models_df <- melt(pltCI_models_df, id.vars="id")
for (row_ix in 1:nrow(mltdCI_models_df)) {
for (type in c("Upper", "Lower")) {
if (length(var_components <- unlist(strsplit(
as.character(mltdCI_models_df[row_ix, "variable"]), type))) > 1) {
#print(sprintf("row_ix:%d; type:%s; ", row_ix, type))
mltdCI_models_df[row_ix, "label"] <- var_components[1]
mltdCI_models_df[row_ix, "data"] <-
unlist(strsplit(var_components[2], "[.]"))[2]
mltdCI_models_df[row_ix, "type"] <- type
break
}
}
}
wideCI_models_df <- reshape(subset(mltdCI_models_df, select=-variable),
timevar="type",
idvar=setdiff(names(mltdCI_models_df), c("type", "value", "variable")),
direction="wide")
#print(wideCI_models_df)
mrgdCI_models_df <- merge(wideCI_models_df, mltd_models_df, all.x=TRUE)
#print(mrgdCI_models_df)
# Merge stats back in if CIs don't exist
goback_vars <- c()
for (var in unique(mltd_models_df$label)) {
for (type in unique(mltd_models_df$data)) {
var_type <- paste0(var, ".", type)
# if this data is already present, next
if (var_type %in% unique(paste(mltd_models_df$label, mltd_models_df$data,
sep=".")))
next
#print(sprintf("var_type:%s", var_type))
goback_vars <- c(goback_vars, var_type)
}
}
if (length(goback_vars) > 0) {
mltd_goback_df <- build_statsCI_data(glb_models_df[, c("id", goback_vars)])
mltd_models_df <- rbind(mltd_models_df, mltd_goback_df)
}
# mltd_models_df <- merge(mltd_models_df, glb_models_df[, c("id", "model_method")],
# all.x=TRUE)
png(paste0(glb_out_pfx, "models_bar.png"), width=480*3, height=480*2)
#print(gp <- myplot_bar(mltd_models_df, "id", "value", colorcol_name="model_method") +
print(gp <- myplot_bar(df=mltd_models_df, xcol_name="id", ycol_names="value") +
geom_errorbar(data=mrgdCI_models_df,
mapping=aes(x=mdl_id, ymax=value.Upper, ymin=value.Lower), width=0.5) +
facet_grid(label ~ data, scales="free") +
theme(axis.text.x = element_text(angle = 90,vjust = 0.5)))## Warning: Removed 5 rows containing missing values (geom_errorbar).dev.off()## quartz_off_screen
## 2print(gp)## Warning: Removed 5 rows containing missing values (geom_errorbar).
dsp_models_cols <- c("id",
glbMdlMetricsEval[glbMdlMetricsEval %in% names(glb_models_df)],
grep("opt.", names(glb_models_df), fixed = TRUE, value = TRUE))
# if (glb_is_classification && glb_is_binomial)
# dsp_models_cols <- c(dsp_models_cols, "opt.prob.threshold.OOB")
print(dsp_models_df <- orderBy(get_model_sel_frmla(), glb_models_df)[, dsp_models_cols])## id max.Accuracy.OOB max.AUCROCR.OOB
## 15 RFE.X.glm 0.7879518 0.8700475
## 3 Max.cor.Y.rcv.1X1.glmnet 0.7722892 0.8197236
## 12 Interact.High.cor.Y.glmnet 0.7698795 0.8189000
## 11 Max.cor.Y.rpart 0.7698795 0.7855854
## 13 Low.cor.X.glmnet 0.7686747 0.8727052
## 5 Max.cor.Y.rcv.3X3.glmnet 0.7662651 0.8197470
## 17 Max.cor.Y.rcv.1X1.Interact.glmnet 0.7662651 0.8197470
## 4 Max.cor.Y.rcv.3X1.glmnet 0.7662651 0.8197353
## 6 Max.cor.Y.rcv.3X5.glmnet 0.7650602 0.8196593
## 7 Max.cor.Y.rcv.5X1.glmnet 0.7638554 0.8195600
## 8 Max.cor.Y.rcv.5X3.glmnet 0.7638554 0.8195600
## 9 Max.cor.Y.rcv.5X5.glmnet 0.7638554 0.8193790
## 14 RFE.X.glmnet 0.7614458 0.8721444
## 16 All.X.glmnet 0.7614458 0.8721444
## 10 Max.cor.Y.rcv.1X1.cp.0.rpart 0.7433735 0.7934621
## 1 MFO.myMFO_classfr 0.4614458 0.5000000
## 2 Random.myrandom_classfr 0.4614458 0.4956046
## max.AUCpROC.OOB max.Accuracy.fit opt.prob.threshold.fit
## 15 0.7851882 0.8307135 0.4
## 3 0.7665376 0.7843137 0.4
## 12 0.7551533 0.8026098 0.4
## 11 0.7445254 0.7964073 0.5
## 13 0.7801532 0.8349657 0.5
## 5 0.7603840 0.7954221 0.5
## 17 0.7603840 0.7954221 0.5
## 4 0.7592654 0.7950911 0.5
## 6 0.7603840 0.7962655 0.5
## 7 0.7615026 0.7990899 0.5
## 8 0.7615026 0.7964364 0.5
## 9 0.7641135 0.7970687 0.5
## 14 0.7835089 0.8303886 0.4
## 16 0.7835089 0.8303886 0.4
## 10 0.7504950 0.8323529 0.4
## 1 0.5000000 0.4627451 0.4
## 2 0.5064252 0.4627451 0.4
## opt.prob.threshold.OOB
## 15 0.5
## 3 0.4
## 12 0.3
## 11 0.3
## 13 0.3
## 5 0.4
## 17 0.4
## 4 0.4
## 6 0.4
## 7 0.4
## 8 0.4
## 9 0.4
## 14 0.3
## 16 0.3
## 10 0.3
## 1 0.4
## 2 0.4print(myplot_radar(radar_inp_df = dsp_models_df))## Warning in RColorBrewer::brewer.pal(n, pal): n too large, allowed maximum for palette Set1 is 9
## Returning the palette you asked for with that many colors## Warning: The shape palette can deal with a maximum of 6 discrete values
## because more than 6 becomes difficult to discriminate; you have
## 17. Consider specifying shapes manually if you must have them.## Warning: Removed 77 rows containing missing values (geom_point).## Warning in RColorBrewer::brewer.pal(n, pal): n too large, allowed maximum for palette Set1 is 9
## Returning the palette you asked for with that many colors## Warning: The shape palette can deal with a maximum of 6 discrete values
## because more than 6 becomes difficult to discriminate; you have
## 17. Consider specifying shapes manually if you must have them.
print("Metrics used for model selection:"); print(get_model_sel_frmla())## [1] "Metrics used for model selection:"## ~-max.Accuracy.OOB - max.AUCROCR.OOB - max.AUCpROC.OOB - max.Accuracy.fit -
## opt.prob.threshold.OOB
## <environment: 0x7fc2f049ec30>print(sprintf("Best model id: %s", dsp_models_df[1, "id"]))## [1] "Best model id: RFE.X.glm"glb_get_predictions <- function(df, mdl_id, rsp_var_out, prob_threshold_def=NULL, verbose=FALSE) {
mdl <- glb_models_lst[[mdl_id]]
#rsp_var_out <- paste0(rsp_var_out, mdl_id)
rsp_var_out <- paste0(glb_rsp_var, ".predict.")
predct_var_name <- paste0(rsp_var_out, mdl_id)
predct_prob_var_name <- paste0(rsp_var_out, mdl_id, ".prob")
predct_accurate_var_name <- paste0(rsp_var_out, mdl_id, ".accurate")
predct_error_var_name <- paste0(rsp_var_out, mdl_id, ".err")
predct_erabs_var_name <- paste0(rsp_var_out, mdl_id, ".err.abs")
if (glb_is_regression) {
df[, predct_var_name] <- predict(mdl, newdata=df, type="raw")
if (verbose) print(myplot_scatter(df, glb_rsp_var, predct_var_name) +
facet_wrap(reformulate(glb_category_var), scales = "free") +
stat_smooth(method="glm"))
df[, predct_error_var_name] <- df[, predct_var_name] - df[, glb_rsp_var]
if (verbose) print(myplot_scatter(df, predct_var_name, predct_error_var_name) +
#facet_wrap(reformulate(glb_category_var), scales = "free") +
stat_smooth(method="auto"))
if (verbose) print(myplot_scatter(df, glb_rsp_var, predct_error_var_name) +
#facet_wrap(reformulate(glb_category_var), scales = "free") +
stat_smooth(method="glm"))
df[, predct_erabs_var_name] <- abs(df[, predct_error_var_name])
if (verbose) print(head(orderBy(reformulate(c("-", predct_erabs_var_name)), df)))
df[, predct_accurate_var_name] <- (df[, glb_rsp_var] == df[, predct_var_name])
}
if (glb_is_classification && glb_is_binomial) {
prob_threshold <- glb_models_df[glb_models_df$id == mdl_id,
"opt.prob.threshold.OOB"]
if (is.null(prob_threshold) || is.na(prob_threshold)) {
warning("Using default probability threshold: ", prob_threshold_def)
if (is.null(prob_threshold <- prob_threshold_def))
stop("Default probability threshold is NULL")
}
df[, predct_prob_var_name] <- predict(mdl, newdata = df, type = "prob")[, 2]
df[, predct_var_name] <-
factor(levels(df[, glb_rsp_var])[
(df[, predct_prob_var_name] >=
prob_threshold) * 1 + 1], levels(df[, glb_rsp_var]))
# if (verbose) print(myplot_scatter(df, glb_rsp_var, predct_var_name) +
# facet_wrap(reformulate(glb_category_var), scales = "free") +
# stat_smooth(method="glm"))
df[, predct_error_var_name] <- df[, predct_var_name] != df[, glb_rsp_var]
# if (verbose) print(myplot_scatter(df, predct_var_name, predct_error_var_name) +
# #facet_wrap(reformulate(glb_category_var), scales = "free") +
# stat_smooth(method="auto"))
# if (verbose) print(myplot_scatter(df, glb_rsp_var, predct_error_var_name) +
# #facet_wrap(reformulate(glb_category_var), scales = "free") +
# stat_smooth(method="glm"))
# if prediction is a TP (true +ve), measure distance from 1.0
tp <- which((df[, predct_var_name] == df[, glb_rsp_var]) &
(df[, predct_var_name] == levels(df[, glb_rsp_var])[2]))
df[tp, predct_erabs_var_name] <- abs(1 - df[tp, predct_prob_var_name])
#rowIx <- which.max(df[tp, predct_erabs_var_name]); df[tp, c(glb_id_var, glb_rsp_var, predct_var_name, predct_prob_var_name, predct_erabs_var_name)][rowIx, ]
# if prediction is a TN (true -ve), measure distance from 0.0
tn <- which((df[, predct_var_name] == df[, glb_rsp_var]) &
(df[, predct_var_name] == levels(df[, glb_rsp_var])[1]))
df[tn, predct_erabs_var_name] <- abs(0 - df[tn, predct_prob_var_name])
#rowIx <- which.max(df[tn, predct_erabs_var_name]); df[tn, c(glb_id_var, glb_rsp_var, predct_var_name, predct_prob_var_name, predct_erabs_var_name)][rowIx, ]
# if prediction is a FP (flse +ve), measure distance from 0.0
fp <- which((df[, predct_var_name] != df[, glb_rsp_var]) &
(df[, predct_var_name] == levels(df[, glb_rsp_var])[2]))
df[fp, predct_erabs_var_name] <- abs(0 - df[fp, predct_prob_var_name])
#rowIx <- which.max(df[fp, predct_erabs_var_name]); df[fp, c(glb_id_var, glb_rsp_var, predct_var_name, predct_prob_var_name, predct_erabs_var_name)][rowIx, ]
# if prediction is a FN (flse -ve), measure distance from 1.0
fn <- which((df[, predct_var_name] != df[, glb_rsp_var]) &
(df[, predct_var_name] == levels(df[, glb_rsp_var])[1]))
df[fn, predct_erabs_var_name] <- abs(1 - df[fn, predct_prob_var_name])
#rowIx <- which.max(df[fn, predct_erabs_var_name]); df[fn, c(glb_id_var, glb_rsp_var, predct_var_name, predct_prob_var_name, predct_erabs_var_name)][rowIx, ]
if (verbose) print(head(orderBy(reformulate(c("-", predct_erabs_var_name)), df)))
df[, predct_accurate_var_name] <- (df[, glb_rsp_var] == df[, predct_var_name])
}
if (glb_is_classification && !glb_is_binomial) {
df[, predct_var_name] <- predict(mdl, newdata = df, type = "raw")
df[, paste0(predct_var_name, ".prob")] <-
predict(mdl, newdata = df, type = "prob")
stop("Multinomial prediction error calculation needs to be implemented...")
}
return(df)
}
#stop(here"); glb_to_sav(); glb_allobs_df <- sav_allobs_df; glb_trnobs_df <- sav_trnobs_df; glb_fitobs_df <- sav_fitobs_df; glb_OOBobs_df <- sav_OOBobs_df; sav_models_df <- glb_models_df; glb_models_df <- sav_models_df; glb_featsimp_df <- sav_featsimp_df
myget_category_stats <- function(obs_df, mdl_id, label) {
require(dplyr)
require(lazyeval)
predct_var_name <- paste0(glb_rsp_var_out, mdl_id)
predct_error_var_name <- paste0(glb_rsp_var_out, mdl_id, ".err.abs")
if (!predct_var_name %in% names(obs_df))
obs_df <- glb_get_predictions(obs_df, mdl_id, glb_rsp_var_out)
tmp_obs_df <- obs_df %>%
dplyr::select_(glb_category_var, glb_rsp_var, predct_var_name, predct_error_var_name)
#dplyr::rename(startprice.log10.predict.RFE.X.glmnet.err=error_abs_OOB)
names(tmp_obs_df)[length(names(tmp_obs_df))] <- paste0("err.abs.", label)
ret_ctgry_df <- tmp_obs_df %>%
dplyr::group_by_(glb_category_var) %>%
dplyr::summarise_(#interp(~sum(abs(var)), var=as.name(glb_rsp_var)),
interp(~sum(var), var=as.name(paste0("err.abs.", label))),
interp(~mean(var), var=as.name(paste0("err.abs.", label))),
interp(~n()))
names(ret_ctgry_df) <- c(glb_category_var,
#paste0(glb_rsp_var, ".abs.", label, ".sum"),
paste0("err.abs.", label, ".sum"),
paste0("err.abs.", label, ".mean"),
paste0(".n.", label))
ret_ctgry_df <- dplyr::ungroup(ret_ctgry_df)
#colSums(ret_ctgry_df[, -grep(glb_category_var, names(ret_ctgry_df))])
return(ret_ctgry_df)
}
#print(colSums((ctgry_df <- myget_category_stats(obs_df=glb_fitobs_df, mdl_id="", label="fit"))[, -grep(glb_category_var, names(ctgry_df))]))
if (!is.null(glb_mdl_ensemble)) {
mdl_id_pfx <- "Ensemble"
if (#(glb_is_regression) |
((glb_is_classification) & (!glb_is_binomial)))
stop("Ensemble models not implemented yet for multinomial classification")
if (glb_mdl_ensemble == "auto") {
mdl_id_pfx <- paste0(mdl_id_pfx, ".auto")
tmp_models_df <- orderBy(get_model_sel_frmla(), glb_models_df)
row.names(tmp_models_df) <- tmp_models_df$id
# mdl_threshold_pos <- min(which(tmp_models_df$id %in%
# c("MFO.myMFO_classfr", "Baseline.mybaseln_classfr"))) - 1
mdl_threshold_pos <-
min(which(grepl("MFO|Random|Baseline", tmp_models_df$id))) - 1
glb_mdl_ensemble <- tmp_models_df$id[1:mdl_threshold_pos]
}
for (mdl_id in glb_mdl_ensemble) {
if (!(mdl_id %in% names(glb_models_lst))) {
warning("Model ", mdl_id, " in glb_model_ensemble not found !")
next
}
glb_fitobs_df <- glb_get_predictions(df = glb_fitobs_df, mdl_id,
glb_rsp_var_out)
glb_OOBobs_df <- glb_get_predictions(df = glb_OOBobs_df, mdl_id,
glb_rsp_var_out)
}
#mdl_id_pfx <- "Ensemble.RFE"; mdlId <- paste0(mdl_id_pfx, ".glmnet")
#glb_mdl_ensemble <- gsub(glb_rsp_var_out, "", grep("RFE\\.X\\.(?!Interact)", row.names(glb_featsimp_df), perl = TRUE, value = TRUE), fixed = TRUE)
#varImp(glb_models_lst[[mdlId]])
#cor_df <- data.frame(cor=cor(glb_fitobs_df[, glb_rsp_var], glb_fitobs_df[, paste(glb_rsp_var_out, glb_mdl_ensemble)], use="pairwise.complete.obs"))
#glb_fitobs_df <- glb_get_predictions(df=glb_fitobs_df, "Ensemble.glmnet", glb_rsp_var_out);print(colSums((ctgry_df <- myget_category_stats(obs_df=glb_fitobs_df, mdl_id="Ensemble.glmnet", label="fit"))[, -grep(glb_category_var, names(ctgry_df))]))
### bid0_sp
# Better than MFO; models.n=28; min.RMSE.fit=0.0521233; err.abs.fit.sum=7.3631895
# old: Top x from auto; models.n= 5; min.RMSE.fit=0.06311047; err.abs.fit.sum=9.5937080
# RFE only ; models.n=16; min.RMSE.fit=0.05148588; err.abs.fit.sum=7.2875091
# RFE subset only ;models.n= 5; min.RMSE.fit=0.06040702; err.abs.fit.sum=9.059088
# RFE subset only ;models.n= 9; min.RMSE.fit=0.05933167; err.abs.fit.sum=8.7421288
# RFE subset only ;models.n=15; min.RMSE.fit=0.0584607; err.abs.fit.sum=8.5902066
# RFE subset only ;models.n=17; min.RMSE.fit=0.05496899; err.abs.fit.sum=8.0170431
# RFE subset only ;models.n=18; min.RMSE.fit=0.05441577; err.abs.fit.sum=7.837223
# RFE subset only ;models.n=16; min.RMSE.fit=0.05441577; err.abs.fit.sum=7.837223
### bid0_sp
### bid1_sp
# "auto"; err.abs.fit.sum=76.699774; min.RMSE.fit=0.2186429
# "RFE.X.*"; err.abs.fit.sum=; min.RMSE.fit=0.221114
### bid1_sp
indep_vars <- paste(glb_rsp_var_out, glb_mdl_ensemble, sep = "")
if (glb_is_classification)
indep_vars <- paste(indep_vars, ".prob", sep = "")
# Some models in glb_mdl_ensemble might not be fitted e.g. RFE.X.Interact
indep_vars <- intersect(indep_vars, names(glb_fitobs_df))
# indep_vars <- grep(glb_rsp_var_out, names(glb_fitobs_df), fixed=TRUE, value=TRUE)
# if (glb_is_regression)
# indep_vars <- indep_vars[!grepl("(err\\.abs|accurate)$", indep_vars)]
# if (glb_is_classification && glb_is_binomial)
# indep_vars <- grep("prob$", indep_vars, value=TRUE) else
# indep_vars <- indep_vars[!grepl("err$", indep_vars)]
#rfe_fit_ens_results <- myrun_rfe(glb_fitobs_df, indep_vars)
for (method in c("glm", "glmnet")) {
for (trainControlMethod in
c("boot", "boot632", "cv", "repeatedcv"
#, "LOOCV" # tuneLength * nrow(fitDF)
, "LGOCV", "adaptive_cv"
#, "adaptive_boot" #error: adaptive$min should be less than 3
#, "adaptive_LGOCV" #error: adaptive$min should be less than 3
)) {
#sav_models_df <- glb_models_df; all.equal(sav_models_df, glb_models_df)
#glb_models_df <- sav_models_df; print(glb_models_df$id)
if ((method == "glm") && (trainControlMethod != "repeatedcv"))
# glm used only to identify outliers
next
ret_lst <- myfit_mdl(
mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = paste0(mdl_id_pfx, ".", trainControlMethod),
type = glb_model_type, tune.df = NULL,
trainControl.method = trainControlMethod,
trainControl.number = glb_rcv_n_folds,
trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = method)),
indep_vars = indep_vars, rsp_var = glb_rsp_var,
fit_df = glb_fitobs_df, OOB_df = glb_OOBobs_df)
}
}
dsp_models_df <- get_dsp_models_df()
}
if (is.null(glb_sel_mdl_id))
glb_sel_mdl_id <- dsp_models_df[1, "id"] else
print(sprintf("User specified selection: %s", glb_sel_mdl_id)) ## [1] "User specified selection: All.X.glmnet"myprint_mdl(glb_sel_mdl <- glb_models_lst[[glb_sel_mdl_id]])
## Length Class Mode
## a0 100 -none- numeric
## beta 9300 dgCMatrix S4
## df 100 -none- numeric
## dim 2 -none- numeric
## lambda 100 -none- numeric
## dev.ratio 100 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 93 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept)
## 5.9298599950
## .rnorm
## -0.0367487387
## D.chrs.n.log
## -0.2852958123
## D.chrs.pnct11.n.log
## 0.4714450805
## D.chrs.pnct13.n.log
## -0.1653784579
## D.chrs.uppr.n.log
## -0.0116026565
## D.ratio.weight.sum.wrds.n
## -0.0593199934
## D.ratio.wrds.stop.n.wrds.n
## -2.9071388332
## D.terms.post.stem.n.log
## -0.2864766681
## D.terms.post.stop.n.log
## -0.9265313808
## D.weight.post.stem.sum
## -0.1209650771
## D.weight.sum
## -0.1180951219
## D.weight.sum.stem.stop.Ratio
## -2.9013605803
## D.wrds.n.log
## 1.2607015439
## D.wrds.stop.n.log
## 0.1152103538
## D.wrds.unq.n.log
## -0.1770678274
## biddable
## 2.5343114368
## cellular.fctrUnknown
## -0.7922522457
## color.fctrGold
## 0.0871372337
## color.fctrSpace Gray
## -0.3474018561
## color.fctrUnknown
## 0.3560862355
## color.fctrWhite
## -0.2253657549
## condition.fctrFor parts or not working
## -0.7365978083
## condition.fctrManufacturer refurbished
## -1.1969859100
## condition.fctrNew
## 0.0297417566
## condition.fctrNew other (see details)
## 0.7229925582
## condition.fctrSeller refurbished
## -1.0271638293
## prdl.my.fctriPad1
## -0.6856562063
## prdl.my.fctriPad2
## -0.4622200902
## prdl.my.fctriPad3
## 0.1837795731
## prdl.my.fctriPad4
## 1.4053112779
## prdl.my.fctriPadAir
## 1.5486082406
## prdl.my.fctriPadAir2
## 1.9064724435
## prdl.my.fctriPadmini
## -0.4755225796
## prdl.my.fctriPadmini2
## 1.3324979842
## prdl.my.fctriPadmini3
## 0.7573021470
## spdiff.cut.fctr(-100,-10]
## 2.3577551693
## spdiff.cut.fctr(-10,-1]
## 4.1547129846
## spdiff.cut.fctr(-1,0]
## 5.3348860714
## spdiff.cut.fctr(0,1]
## 4.4072812301
## spdiff.cut.fctr(1,10]
## 4.4073453219
## spdiff.cut.fctr(10,100]
## 3.8533894456
## spdiff.cut.fctr(100,1e+03]
## 1.0918155713
## sprice.d20nexp
## -1.5657841541
## sprice.log10
## -0.5769819282
## sprice.root2
## -0.1085201562
## startprice.dcm1.is9
## -0.4894970169
## startprice.dcm2.is9
## -0.0838883551
## startprice.dgt1.is9
## 0.0254181773
## startprice.dgt2.is9
## 0.1157259996
## storage.fctr16
## -0.1300313626
## storage.fctr32
## 0.0005589105
## storage.fctr64
## 0.2180851783
## storage.fctrUnknown
## 1.7336441556
## cellular.fctr0:carrier.fctrNone
## 0.1706810167
## cellular.fctr1:carrier.fctrOther
## 2.6072585524
## cellular.fctr1:carrier.fctrSprint
## 1.0270604082
## cellular.fctr1:carrier.fctrT-Mobile
## -0.2821763037
## cellular.fctr1:carrier.fctrUnknown
## -0.0179172973
## cellular.fctrUnknown:carrier.fctrUnknown
## -0.7883348835
## cellular.fctr1:carrier.fctrVerizon
## 0.3156224380
## prdl.my.fctrUnknown:.clusterid.fctr2
## 0.8932344300
## prdl.my.fctriPad1:.clusterid.fctr2
## -0.7404516345
## prdl.my.fctriPad3:.clusterid.fctr2
## 0.2701330403
## prdl.my.fctriPad4:.clusterid.fctr2
## -1.4552193885
## prdl.my.fctriPadAir:.clusterid.fctr2
## -0.3279730199
## prdl.my.fctriPadmini2:.clusterid.fctr2
## -0.6287767988
## prdl.my.fctrUnknown:.clusterid.fctr3
## -0.7598398151
## prdl.my.fctriPad1:.clusterid.fctr3
## -1.1337789530
## prdl.my.fctriPad4:.clusterid.fctr3
## -5.9466674332
## [1] "max lambda < lambdaOpt:"
## [1] "Feats mismatch between coefs_left & rght:"
## [1] "(Intercept)"
## [2] ".rnorm"
## [3] "D.chrs.n.log"
## [4] "D.chrs.pnct11.n.log"
## [5] "D.chrs.pnct13.n.log"
## [6] "D.chrs.uppr.n.log"
## [7] "D.ratio.weight.sum.wrds.n"
## [8] "D.ratio.wrds.stop.n.wrds.n"
## [9] "D.terms.post.stem.n.log"
## [10] "D.terms.post.stop.n.log"
## [11] "D.weight.post.stem.sum"
## [12] "D.weight.post.stop.sum"
## [13] "D.weight.sum"
## [14] "D.weight.sum.stem.stop.Ratio"
## [15] "D.wrds.n.log"
## [16] "D.wrds.stop.n.log"
## [17] "D.wrds.unq.n.log"
## [18] "biddable"
## [19] "cellular.fctr1"
## [20] "cellular.fctrUnknown"
## [21] "color.fctrGold"
## [22] "color.fctrSpace Gray"
## [23] "color.fctrUnknown"
## [24] "color.fctrWhite"
## [25] "condition.fctrFor parts or not working"
## [26] "condition.fctrManufacturer refurbished"
## [27] "condition.fctrNew"
## [28] "condition.fctrNew other (see details)"
## [29] "condition.fctrSeller refurbished"
## [30] "prdl.my.fctriPad1"
## [31] "prdl.my.fctriPad2"
## [32] "prdl.my.fctriPad3"
## [33] "prdl.my.fctriPad4"
## [34] "prdl.my.fctriPadAir"
## [35] "prdl.my.fctriPadAir2"
## [36] "prdl.my.fctriPadmini"
## [37] "prdl.my.fctriPadmini2"
## [38] "prdl.my.fctriPadmini3"
## [39] "spdiff.cut.fctr(-100,-10]"
## [40] "spdiff.cut.fctr(-10,-1]"
## [41] "spdiff.cut.fctr(-1,0]"
## [42] "spdiff.cut.fctr(0,1]"
## [43] "spdiff.cut.fctr(1,10]"
## [44] "spdiff.cut.fctr(10,100]"
## [45] "spdiff.cut.fctr(100,1e+03]"
## [46] "sprice.d20nexp"
## [47] "sprice.log10"
## [48] "sprice.root2"
## [49] "startprice.dcm1.is9"
## [50] "startprice.dcm2.is9"
## [51] "startprice.dgt1.is9"
## [52] "startprice.dgt2.is9"
## [53] "storage.fctr16"
## [54] "storage.fctr32"
## [55] "storage.fctr64"
## [56] "storage.fctrUnknown"
## [57] "cellular.fctr0:carrier.fctrNone"
## [58] "cellular.fctr1:carrier.fctrNone"
## [59] "cellular.fctrUnknown:carrier.fctrNone"
## [60] "cellular.fctr0:carrier.fctrOther"
## [61] "cellular.fctr1:carrier.fctrOther"
## [62] "cellular.fctrUnknown:carrier.fctrOther"
## [63] "cellular.fctr0:carrier.fctrSprint"
## [64] "cellular.fctr1:carrier.fctrSprint"
## [65] "cellular.fctrUnknown:carrier.fctrSprint"
## [66] "cellular.fctr0:carrier.fctrT-Mobile"
## [67] "cellular.fctr1:carrier.fctrT-Mobile"
## [68] "cellular.fctrUnknown:carrier.fctrT-Mobile"
## [69] "cellular.fctr0:carrier.fctrUnknown"
## [70] "cellular.fctr1:carrier.fctrUnknown"
## [71] "cellular.fctrUnknown:carrier.fctrUnknown"
## [72] "cellular.fctr0:carrier.fctrVerizon"
## [73] "cellular.fctr1:carrier.fctrVerizon"
## [74] "cellular.fctrUnknown:carrier.fctrVerizon"
## [75] "prdl.my.fctrUnknown:.clusterid.fctr2"
## [76] "prdl.my.fctriPad1:.clusterid.fctr2"
## [77] "prdl.my.fctriPad2:.clusterid.fctr2"
## [78] "prdl.my.fctriPad3:.clusterid.fctr2"
## [79] "prdl.my.fctriPad4:.clusterid.fctr2"
## [80] "prdl.my.fctriPadAir:.clusterid.fctr2"
## [81] "prdl.my.fctriPadAir2:.clusterid.fctr2"
## [82] "prdl.my.fctriPadmini:.clusterid.fctr2"
## [83] "prdl.my.fctriPadmini2:.clusterid.fctr2"
## [84] "prdl.my.fctriPadmini3:.clusterid.fctr2"
## [85] "prdl.my.fctrUnknown:.clusterid.fctr3"
## [86] "prdl.my.fctriPad1:.clusterid.fctr3"
## [87] "prdl.my.fctriPad2:.clusterid.fctr3"
## [88] "prdl.my.fctriPad3:.clusterid.fctr3"
## [89] "prdl.my.fctriPad4:.clusterid.fctr3"
## [90] "prdl.my.fctriPadAir:.clusterid.fctr3"
## [91] "prdl.my.fctriPadAir2:.clusterid.fctr3"
## [92] "prdl.my.fctriPadmini:.clusterid.fctr3"
## [93] "prdl.my.fctriPadmini2:.clusterid.fctr3"
## [94] "prdl.my.fctriPadmini3:.clusterid.fctr3"## [1] TRUE#stop(here"); glb_to_sav()
# From here to save(), this should all be in one function
# these are executed in the same seq twice more:
# fit.data.training & predict.data.new chunks
print(sprintf("%s fit prediction diagnostics:", glb_sel_mdl_id))## [1] "All.X.glmnet fit prediction diagnostics:"glb_fitobs_df <- glb_get_predictions(df=glb_fitobs_df, mdl_id=glb_sel_mdl_id,
rsp_var_out=glb_rsp_var_out)
print(sprintf("%s OOB prediction diagnostics:", glb_sel_mdl_id))## [1] "All.X.glmnet OOB prediction diagnostics:"glb_OOBobs_df <- glb_get_predictions(df = glb_OOBobs_df, mdl_id = glb_sel_mdl_id,
rsp_var_out = glb_rsp_var_out)
glb_featsimp_df <-
myget_feats_importance(mdl=glb_sel_mdl, featsimp_df=NULL)
glb_featsimp_df[, paste0(glb_sel_mdl_id, ".importance")] <- glb_featsimp_df$importance
#mdl_id <-"RFE.X.glmnet"; glb_featsimp_df <- myget_feats_importance(glb_models_lst[[mdl_id]], glb_featsimp_df); glb_featsimp_df[, paste0(mdl_id, ".importance")] <- glb_featsimp_df$importance; print(glb_featsimp_df)
#print(head(sbst_featsimp_df <- subset(glb_featsimp_df, is.na(RFE.X.glmnet.importance) | (abs(RFE.X.YeoJohnson.glmnet.importance - RFE.X.glmnet.importance) > 0.0001), select=-importance)))
#print(orderBy(~ -cor.y.abs, subset(glb_feats_df, id %in% c(row.names(sbst_featsimp_df), "startprice.dcm1.is9", "D.weight.post.stop.sum"))))
print(glb_featsimp_df)## importance
## spdiff.cut.fctr(-1,0] 100.00000
## spdiff.cut.fctr(1,10] 91.77825
## spdiff.cut.fctr(0,1] 91.77769
## spdiff.cut.fctr(-10,-1] 89.53891
## spdiff.cut.fctr(10,100] 86.86797
## cellular.fctr1:carrier.fctrOther 75.82223
## biddable 75.17563
## spdiff.cut.fctr(-100,-10] 73.61063
## prdl.my.fctriPadAir2 69.61045
## storage.fctrUnknown 68.07849
## prdl.my.fctriPadAir 66.43833
## prdl.my.fctriPad4 65.16814
## prdl.my.fctriPadmini2 64.52272
## D.wrds.n.log 63.88632
## spdiff.cut.fctr(100,1e+03] 62.38931
## cellular.fctr1:carrier.fctrSprint 61.81532
## prdl.my.fctrUnknown:.clusterid.fctr2 60.62908
## prdl.my.fctriPadmini3 59.42417
## condition.fctrNew other (see details) 59.12005
## D.chrs.pnct11.n.log 56.89033
## color.fctrUnknown 55.86778
## cellular.fctr1:carrier.fctrVerizon 55.50911
## prdl.my.fctriPad3:.clusterid.fctr2 55.10589
## storage.fctr64 54.64454
## prdl.my.fctriPad3 54.34045
## cellular.fctr0:carrier.fctrNone 54.22434
## startprice.dgt2.is9 53.73722
## D.wrds.stop.n.log 53.73265
## color.fctrGold 53.48381
## condition.fctrNew 52.97506
## startprice.dgt1.is9 52.93673
## storage.fctr32 52.71638
## D.weight.post.stop.sum 52.71142
## cellular.fctr1 52.71142
## cellular.fctr1:carrier.fctrNone 52.71142
## cellular.fctrUnknown:carrier.fctrNone 52.71142
## cellular.fctr0:carrier.fctrOther 52.71142
## cellular.fctrUnknown:carrier.fctrOther 52.71142
## cellular.fctr0:carrier.fctrSprint 52.71142
## cellular.fctrUnknown:carrier.fctrSprint 52.71142
## cellular.fctr0:carrier.fctrT-Mobile 52.71142
## cellular.fctrUnknown:carrier.fctrT-Mobile 52.71142
## cellular.fctr0:carrier.fctrUnknown 52.71142
## cellular.fctr0:carrier.fctrVerizon 52.71142
## cellular.fctrUnknown:carrier.fctrVerizon 52.71142
## prdl.my.fctriPad2:.clusterid.fctr2 52.71142
## prdl.my.fctriPadAir2:.clusterid.fctr2 52.71142
## prdl.my.fctriPadmini:.clusterid.fctr2 52.71142
## prdl.my.fctriPadmini3:.clusterid.fctr2 52.71142
## prdl.my.fctriPad2:.clusterid.fctr3 52.71142
## prdl.my.fctriPad3:.clusterid.fctr3 52.71142
## prdl.my.fctriPadAir:.clusterid.fctr3 52.71142
## prdl.my.fctriPadAir2:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini2:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini3:.clusterid.fctr3 52.71142
## D.chrs.uppr.n.log 52.60858
## cellular.fctr1:carrier.fctrUnknown 52.55260
## .rnorm 52.38568
## D.ratio.weight.sum.wrds.n 52.18561
## startprice.dcm2.is9 51.96783
## sprice.root2 51.74950
## D.weight.sum 51.66462
## D.weight.post.stem.sum 51.63919
## storage.fctr16 51.55882
## D.chrs.pnct13.n.log 51.24550
## D.wrds.unq.n.log 51.14189
## color.fctrWhite 50.71378
## cellular.fctr1:carrier.fctrT-Mobile 50.21020
## D.chrs.n.log 50.18255
## D.terms.post.stem.n.log 50.17209
## prdl.my.fctriPadAir:.clusterid.fctr2 49.80426
## color.fctrSpace Gray 49.63204
## prdl.my.fctriPad2 48.61429
## prdl.my.fctriPadmini 48.49638
## startprice.dcm1.is9 48.37251
## sprice.log10 47.59704
## prdl.my.fctriPadmini2:.clusterid.fctr2 47.13793
## prdl.my.fctriPad1 46.63375
## condition.fctrFor parts or not working 46.18220
## prdl.my.fctriPad1:.clusterid.fctr2 46.14804
## prdl.my.fctrUnknown:.clusterid.fctr3 45.97618
## cellular.fctrUnknown:carrier.fctrUnknown 45.72360
## cellular.fctrUnknown 45.68888
## D.terms.post.stop.n.log 44.49862
## condition.fctrSeller refurbished 43.60662
## prdl.my.fctriPad1:.clusterid.fctr3 42.66158
## condition.fctrManufacturer refurbished 42.10131
## prdl.my.fctriPad4:.clusterid.fctr2 39.81232
## sprice.d20nexp 38.83227
## D.weight.sum.stem.stop.Ratio 26.99368
## D.ratio.wrds.stop.n.wrds.n 26.94246
## prdl.my.fctriPad4:.clusterid.fctr3 0.00000
## All.X.glmnet.importance
## spdiff.cut.fctr(-1,0] 100.00000
## spdiff.cut.fctr(1,10] 91.77825
## spdiff.cut.fctr(0,1] 91.77769
## spdiff.cut.fctr(-10,-1] 89.53891
## spdiff.cut.fctr(10,100] 86.86797
## cellular.fctr1:carrier.fctrOther 75.82223
## biddable 75.17563
## spdiff.cut.fctr(-100,-10] 73.61063
## prdl.my.fctriPadAir2 69.61045
## storage.fctrUnknown 68.07849
## prdl.my.fctriPadAir 66.43833
## prdl.my.fctriPad4 65.16814
## prdl.my.fctriPadmini2 64.52272
## D.wrds.n.log 63.88632
## spdiff.cut.fctr(100,1e+03] 62.38931
## cellular.fctr1:carrier.fctrSprint 61.81532
## prdl.my.fctrUnknown:.clusterid.fctr2 60.62908
## prdl.my.fctriPadmini3 59.42417
## condition.fctrNew other (see details) 59.12005
## D.chrs.pnct11.n.log 56.89033
## color.fctrUnknown 55.86778
## cellular.fctr1:carrier.fctrVerizon 55.50911
## prdl.my.fctriPad3:.clusterid.fctr2 55.10589
## storage.fctr64 54.64454
## prdl.my.fctriPad3 54.34045
## cellular.fctr0:carrier.fctrNone 54.22434
## startprice.dgt2.is9 53.73722
## D.wrds.stop.n.log 53.73265
## color.fctrGold 53.48381
## condition.fctrNew 52.97506
## startprice.dgt1.is9 52.93673
## storage.fctr32 52.71638
## D.weight.post.stop.sum 52.71142
## cellular.fctr1 52.71142
## cellular.fctr1:carrier.fctrNone 52.71142
## cellular.fctrUnknown:carrier.fctrNone 52.71142
## cellular.fctr0:carrier.fctrOther 52.71142
## cellular.fctrUnknown:carrier.fctrOther 52.71142
## cellular.fctr0:carrier.fctrSprint 52.71142
## cellular.fctrUnknown:carrier.fctrSprint 52.71142
## cellular.fctr0:carrier.fctrT-Mobile 52.71142
## cellular.fctrUnknown:carrier.fctrT-Mobile 52.71142
## cellular.fctr0:carrier.fctrUnknown 52.71142
## cellular.fctr0:carrier.fctrVerizon 52.71142
## cellular.fctrUnknown:carrier.fctrVerizon 52.71142
## prdl.my.fctriPad2:.clusterid.fctr2 52.71142
## prdl.my.fctriPadAir2:.clusterid.fctr2 52.71142
## prdl.my.fctriPadmini:.clusterid.fctr2 52.71142
## prdl.my.fctriPadmini3:.clusterid.fctr2 52.71142
## prdl.my.fctriPad2:.clusterid.fctr3 52.71142
## prdl.my.fctriPad3:.clusterid.fctr3 52.71142
## prdl.my.fctriPadAir:.clusterid.fctr3 52.71142
## prdl.my.fctriPadAir2:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini2:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini3:.clusterid.fctr3 52.71142
## D.chrs.uppr.n.log 52.60858
## cellular.fctr1:carrier.fctrUnknown 52.55260
## .rnorm 52.38568
## D.ratio.weight.sum.wrds.n 52.18561
## startprice.dcm2.is9 51.96783
## sprice.root2 51.74950
## D.weight.sum 51.66462
## D.weight.post.stem.sum 51.63919
## storage.fctr16 51.55882
## D.chrs.pnct13.n.log 51.24550
## D.wrds.unq.n.log 51.14189
## color.fctrWhite 50.71378
## cellular.fctr1:carrier.fctrT-Mobile 50.21020
## D.chrs.n.log 50.18255
## D.terms.post.stem.n.log 50.17209
## prdl.my.fctriPadAir:.clusterid.fctr2 49.80426
## color.fctrSpace Gray 49.63204
## prdl.my.fctriPad2 48.61429
## prdl.my.fctriPadmini 48.49638
## startprice.dcm1.is9 48.37251
## sprice.log10 47.59704
## prdl.my.fctriPadmini2:.clusterid.fctr2 47.13793
## prdl.my.fctriPad1 46.63375
## condition.fctrFor parts or not working 46.18220
## prdl.my.fctriPad1:.clusterid.fctr2 46.14804
## prdl.my.fctrUnknown:.clusterid.fctr3 45.97618
## cellular.fctrUnknown:carrier.fctrUnknown 45.72360
## cellular.fctrUnknown 45.68888
## D.terms.post.stop.n.log 44.49862
## condition.fctrSeller refurbished 43.60662
## prdl.my.fctriPad1:.clusterid.fctr3 42.66158
## condition.fctrManufacturer refurbished 42.10131
## prdl.my.fctriPad4:.clusterid.fctr2 39.81232
## sprice.d20nexp 38.83227
## D.weight.sum.stem.stop.Ratio 26.99368
## D.ratio.wrds.stop.n.wrds.n 26.94246
## prdl.my.fctriPad4:.clusterid.fctr3 0.00000# Used again in fit.data.training & predict.data.new chunks
glb_analytics_diag_plots <- function(obs_df, mdl_id, prob_threshold=NULL) {
if (!is.null(featsimp_df <- glb_featsimp_df)) {
featsimp_df$feat <- gsub("`(.*?)`", "\\1", row.names(featsimp_df))
featsimp_df$feat.interact <- gsub("(.*?):(.*)", "\\2", featsimp_df$feat)
featsimp_df$feat <- gsub("(.*?):(.*)", "\\1", featsimp_df$feat)
featsimp_df$feat.interact <- ifelse(featsimp_df$feat.interact == featsimp_df$feat,
NA, featsimp_df$feat.interact)
featsimp_df$feat <- gsub("(.*?)\\.fctr(.*)", "\\1\\.fctr", featsimp_df$feat)
featsimp_df$feat.interact <- gsub("(.*?)\\.fctr(.*)", "\\1\\.fctr", featsimp_df$feat.interact)
featsimp_df <- orderBy(~ -importance.max, summaryBy(importance ~ feat + feat.interact,
data=featsimp_df, FUN=max))
#rex_str=":(.*)"; txt_vctr=tail(featsimp_df$feat); ret_lst <- regexec(rex_str, txt_vctr); ret_lst <- regmatches(txt_vctr, ret_lst); ret_vctr <- sapply(1:length(ret_lst), function(pos_ix) ifelse(length(ret_lst[[pos_ix]]) > 0, ret_lst[[pos_ix]], "")); print(ret_vctr <- ret_vctr[ret_vctr != ""])
featsimp_df <- subset(featsimp_df, !is.na(importance.max))
if (nrow(featsimp_df) > 5) {
warning("Limiting important feature scatter plots to 5 out of ", nrow(featsimp_df))
featsimp_df <- head(featsimp_df, 5)
}
# if (!all(is.na(featsimp_df$feat.interact)))
# stop("not implemented yet")
rsp_var_out <- paste0(glb_rsp_var_out, mdl_id)
for (var in featsimp_df$feat) {
plot_df <- melt(obs_df, id.vars=var,
measure.vars=c(glb_rsp_var, rsp_var_out))
# if (var == "<feat_name>") print(myplot_scatter(plot_df, var, "value",
# facet_colcol_name="variable") +
# geom_vline(xintercept=<divider_val>, linetype="dotted")) else
print(myplot_scatter(plot_df, var, "value", colorcol_name="variable",
facet_colcol_name="variable", jitter=TRUE) +
guides(color=FALSE))
}
}
if (glb_is_regression) {
if (is.null(featsimp_df) || (nrow(featsimp_df) == 0))
warning("No important features in glb_fin_mdl") else
print(myplot_prediction_regression(df=obs_df,
feat_x=ifelse(nrow(featsimp_df) > 1, featsimp_df$feat[2],
".rownames"),
feat_y=featsimp_df$feat[1],
rsp_var=glb_rsp_var, rsp_var_out=rsp_var_out,
id_vars=glb_id_var)
# + facet_wrap(reformulate(featsimp_df$feat[2])) # if [1 or 2] is a factor
# + geom_point(aes_string(color="<col_name>.fctr")) # to color the plot
)
}
if (glb_is_classification) {
if (is.null(featsimp_df) || (nrow(featsimp_df) == 0))
warning("No features in selected model are statistically important")
else print(myplot_prediction_classification(df=obs_df,
feat_x=ifelse(nrow(featsimp_df) > 1, featsimp_df$feat[2],
".rownames"),
feat_y=featsimp_df$feat[1],
rsp_var=glb_rsp_var,
rsp_var_out=rsp_var_out,
id_vars=glb_id_var,
prob_threshold=prob_threshold)
# + geom_hline(yintercept=<divider_val>, linetype = "dotted")
)
}
}
if (glb_is_classification && glb_is_binomial)
glb_analytics_diag_plots(obs_df = glb_OOBobs_df, mdl_id = glb_sel_mdl_id,
prob_threshold = glb_models_df[glb_models_df$id == glb_sel_mdl_id,
"opt.prob.threshold.OOB"]) else
glb_analytics_diag_plots(obs_df=glb_OOBobs_df, mdl_id=glb_sel_mdl_id) ## Warning in glb_analytics_diag_plots(obs_df = glb_OOBobs_df, mdl_id =
## glb_sel_mdl_id, : Limiting important feature scatter plots to 5 out of 32
## [1] "Min/Max Boundaries: "
## [1] UniqueID
## [2] sold.fctr
## [3] sold.fctr.predict.All.X.glmnet.prob
## [4] sold.fctr.predict.All.X.glmnet
## [5] sold.fctr.predict.All.X.glmnet.err
## [6] sold.fctr.predict.All.X.glmnet.err.abs
## [7] sold.fctr.predict.All.X.glmnet.accurate
## [8] sold.fctr.predict.All.X.glmnet.error
## [9] .label
## <0 rows> (or 0-length row.names)
## [1] "Inaccurate: "
## UniqueID sold.fctr sold.fctr.predict.All.X.glmnet.prob
## 1696 11705 Y 0.003580620
## 1807 11817 Y 0.008680448
## 1506 11514 Y 0.022355332
## 958 10961 Y 0.023805316
## 457 10458 Y 0.031964256
## 332 10332 Y 0.037172380
## sold.fctr.predict.All.X.glmnet sold.fctr.predict.All.X.glmnet.err
## 1696 N TRUE
## 1807 N TRUE
## 1506 N TRUE
## 958 N TRUE
## 457 N TRUE
## 332 N TRUE
## sold.fctr.predict.All.X.glmnet.err.abs
## 1696 0.9964194
## 1807 0.9913196
## 1506 0.9776447
## 958 0.9761947
## 457 0.9680357
## 332 0.9628276
## sold.fctr.predict.All.X.glmnet.accurate
## 1696 FALSE
## 1807 FALSE
## 1506 FALSE
## 958 FALSE
## 457 FALSE
## 332 FALSE
## sold.fctr.predict.All.X.glmnet.error
## 1696 -0.2964194
## 1807 -0.2913196
## 1506 -0.2776447
## 958 -0.2761947
## 457 -0.2680357
## 332 -0.2628276
## UniqueID sold.fctr sold.fctr.predict.All.X.glmnet.prob
## 1696 11705 Y 0.00358062
## 45 10045 Y 0.26431208
## 308 10308 N 0.34278981
## 838 10841 N 0.52344726
## 511 10512 N 0.66487561
## 580 10581 N 0.89966694
## sold.fctr.predict.All.X.glmnet sold.fctr.predict.All.X.glmnet.err
## 1696 N TRUE
## 45 N TRUE
## 308 Y TRUE
## 838 Y TRUE
## 511 Y TRUE
## 580 Y TRUE
## sold.fctr.predict.All.X.glmnet.err.abs
## 1696 0.9964194
## 45 0.7356879
## 308 0.3427898
## 838 0.5234473
## 511 0.6648756
## 580 0.8996669
## sold.fctr.predict.All.X.glmnet.accurate
## 1696 FALSE
## 45 FALSE
## 308 FALSE
## 838 FALSE
## 511 FALSE
## 580 FALSE
## sold.fctr.predict.All.X.glmnet.error
## 1696 -0.29641938
## 45 -0.03568792
## 308 0.04278981
## 838 0.22344726
## 511 0.36487561
## 580 0.59966694
## UniqueID sold.fctr sold.fctr.predict.All.X.glmnet.prob
## 412 10413 N 0.9743568
## 103 10103 N 0.9814365
## 199 10199 N 0.9869369
## 1464 11471 N 0.9903548
## 490 10491 N 0.9924276
## 1384 11391 N 0.9963958
## sold.fctr.predict.All.X.glmnet sold.fctr.predict.All.X.glmnet.err
## 412 Y TRUE
## 103 Y TRUE
## 199 Y TRUE
## 1464 Y TRUE
## 490 Y TRUE
## 1384 Y TRUE
## sold.fctr.predict.All.X.glmnet.err.abs
## 412 0.9743568
## 103 0.9814365
## 199 0.9869369
## 1464 0.9903548
## 490 0.9924276
## 1384 0.9963958
## sold.fctr.predict.All.X.glmnet.accurate
## 412 FALSE
## 103 FALSE
## 199 FALSE
## 1464 FALSE
## 490 FALSE
## 1384 FALSE
## sold.fctr.predict.All.X.glmnet.error
## 412 0.6743568
## 103 0.6814365
## 199 0.6869369
## 1464 0.6903548
## 490 0.6924276
## 1384 0.6963958
glb_ctgry_df <- merge(glb_ctgry_df,
myget_category_stats(obs_df = glb_fitobs_df, mdl_id = glb_sel_mdl_id, label = "fit"),
by = glb_category_var, all = TRUE)
row.names(glb_ctgry_df) <- glb_ctgry_df[, glb_category_var]
glb_ctgry_df <- merge(glb_ctgry_df,
myget_category_stats(obs_df=glb_OOBobs_df, mdl_id=glb_sel_mdl_id, label="OOB"),
#by=glb_category_var, all=TRUE) glb_ctgry-df already contains .n.OOB ?
all=TRUE)
row.names(glb_ctgry_df) <- glb_ctgry_df[, glb_category_var]
if (any(grepl("OOB", glbMdlMetricsEval)))
print(orderBy(~-err.abs.OOB.mean, glb_ctgry_df)) else
print(orderBy(~-err.abs.fit.mean, glb_ctgry_df))## prdl.my.fctr .n.OOB .n.Fit .n.Tst .freqRatio.Fit .freqRatio.OOB
## iPad1 iPad1 95 130 88 0.12745098 0.11445783
## iPadmini2 iPadmini2 53 54 56 0.05294118 0.06385542
## Unknown Unknown 96 108 92 0.10588235 0.11566265
## iPadmini iPadmini 123 154 111 0.15098039 0.14819277
## iPadAir2 iPadAir2 69 102 62 0.10000000 0.08313253
## iPad2 iPad2 142 144 154 0.14117647 0.17108434
## iPad3 iPad3 61 92 55 0.09019608 0.07349398
## iPadmini3 iPadmini3 36 54 38 0.05294118 0.04337349
## iPadAir iPadAir 82 98 74 0.09607843 0.09879518
## iPad4 iPad4 73 84 68 0.08235294 0.08795181
## .freqRatio.Tst err.abs.fit.sum err.abs.fit.mean .n.fit
## iPad1 0.11027569 23.91786 0.1839835 130
## iPadmini2 0.07017544 14.19476 0.2628659 54
## Unknown 0.11528822 26.52091 0.2455640 108
## iPadmini 0.13909774 30.21777 0.1962193 154
## iPadAir2 0.07769424 21.24831 0.2083168 102
## iPad2 0.19298246 31.86069 0.2212548 144
## iPad3 0.06892231 21.85916 0.2375996 92
## iPadmini3 0.04761905 10.64581 0.1971447 54
## iPadAir 0.09273183 13.80524 0.1408698 98
## iPad4 0.08521303 10.53643 0.1254337 84
## err.abs.OOB.sum err.abs.OOB.mean
## iPad1 31.738295 0.3340873
## iPadmini2 15.149743 0.2858442
## Unknown 25.545260 0.2660965
## iPadmini 32.704594 0.2658910
## iPadAir2 17.785002 0.2577536
## iPad2 35.143654 0.2474905
## iPad3 14.689955 0.2408189
## iPadmini3 8.566058 0.2379461
## iPadAir 15.249371 0.1859679
## iPad4 13.177556 0.1805145print(colSums(glb_ctgry_df[, -grep(glb_category_var, names(glb_ctgry_df))]))## .n.OOB .n.Fit .n.Tst .freqRatio.Fit
## 830.000000 1020.000000 798.000000 1.000000
## .freqRatio.OOB .freqRatio.Tst err.abs.fit.sum err.abs.fit.mean
## 1.000000 1.000000 204.806955 2.019252
## .n.fit err.abs.OOB.sum err.abs.OOB.mean
## 1020.000000 209.749487 2.502411write.csv(glb_OOBobs_df[, c(glb_id_var,
grep(glb_rsp_var, names(glb_OOBobs_df), fixed=TRUE, value=TRUE))],
paste0(gsub(".", "_", paste0(glb_out_pfx, glb_sel_mdl_id), fixed=TRUE),
"_OOBobs.csv"), row.names=FALSE)
glb_chunks_df <- myadd_chunk(glb_chunks_df, "fit.models", major.inc=FALSE)## label step_major step_minor label_minor bgn end elapsed
## 12 fit.models 6 2 2 477.215 498.302 21.087
## 13 fit.models 6 3 3 498.303 NA NA# if (sum(is.na(glb_allobs_df$D.P.http)) > 0)
# stop("fit.models_3: Why is this happening ?")
#stop(here"); glb_to_sav()
sync_glb_obs_df <- function() {
# Merge or cbind ?
for (col in setdiff(names(glb_fitobs_df), names(glb_trnobs_df)))
glb_trnobs_df[glb_trnobs_df$.lcn == "Fit", col] <<- glb_fitobs_df[, col]
for (col in setdiff(names(glb_fitobs_df), names(glb_allobs_df)))
glb_allobs_df[glb_allobs_df$.lcn == "Fit", col] <<- glb_fitobs_df[, col]
if (all(is.na(glb_newobs_df[, glb_rsp_var])))
for (col in setdiff(names(glb_OOBobs_df), names(glb_trnobs_df)))
glb_trnobs_df[glb_trnobs_df$.lcn == "OOB", col] <<- glb_OOBobs_df[, col]
for (col in setdiff(names(glb_OOBobs_df), names(glb_allobs_df)))
glb_allobs_df[glb_allobs_df$.lcn == "OOB", col] <<- glb_OOBobs_df[, col]
}
sync_glb_obs_df()
print(setdiff(names(glb_newobs_df), names(glb_allobs_df)))## character(0)if (glb_save_envir)
save(glb_feats_df,
glb_allobs_df, #glb_trnobs_df, glb_fitobs_df, glb_OOBobs_df, glb_newobs_df,
glb_models_df, dsp_models_df, glb_models_lst, glb_sel_mdl, glb_sel_mdl_id,
glb_model_type,
file=paste0(glb_out_pfx, "selmdl_dsk.RData"))
#load(paste0(glb_out_pfx, "selmdl_dsk.RData"))
rm(ret_lst)## Warning in rm(ret_lst): object 'ret_lst' not foundreplay.petrisim(pn=glb_analytics_pn,
replay.trans=(glb_analytics_avl_objs <- c(glb_analytics_avl_objs,
"model.selected")), flip_coord=TRUE)## time trans "bgn " "fit.data.training.all " "predict.data.new " "end "
## 0.0000 multiple enabled transitions: data.training.all data.new model.selected firing: data.training.all
## 1.0000 1 2 1 0 0
## 1.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction firing: data.new
## 2.0000 2 1 1 1 0
## 2.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction data.new.prediction firing: model.selected
## 3.0000 3 0 2 1 0
glb_chunks_df <- myadd_chunk(glb_chunks_df, "fit.data.training", major.inc=TRUE)## label step_major step_minor label_minor bgn end
## 13 fit.models 6 3 3 498.303 504.741
## 14 fit.data.training 7 0 0 504.742 NA
## elapsed
## 13 6.438
## 14 NAStep 7.0: fit data training
#load(paste0(glb_inp_pfx, "dsk.RData"))
#stop(here"); glb_to_sav()
if (!is.null(glb_fin_mdl_id) && (glb_fin_mdl_id %in% names(glb_models_lst))) {
warning("Final model same as user selected model")
glb_fin_mdl <- glb_models_lst[[glb_fin_mdl_id]]
} else
# if (nrow(glb_fitobs_df) + length(glbObsFitOutliers) == nrow(glb_trnobs_df))
if (!all(is.na(glb_newobs_df[, glb_rsp_var])))
{
warning("Final model same as glb_sel_mdl_id")
glb_fin_mdl_id <- paste0("Final.", glb_sel_mdl_id)
glb_fin_mdl <- glb_sel_mdl
glb_models_lst[[glb_fin_mdl_id]] <- glb_fin_mdl
} else {
if (grepl("RFE", glb_sel_mdl_id) ||
(!is.null(glb_mdl_ensemble) && grepl("RFE", glb_mdl_ensemble))) {
indep_vars <- myadjust_interaction_feats(subset(glb_feats_df,
!nzv & (exclude.as.feat != 1))[, "id"])
rfe_trn_results <- myrun_rfe(glb_trnobs_df, indep_vars, glb_rfe_fit_sizes)
if (!isTRUE(all.equal(sort(predictors(rfe_trn_results)),
sort(predictors(rfe_fit_results))))) {
print("Diffs predictors(rfe_trn_results) vs. predictors(rfe_fit_results):")
print(setdiff(predictors(rfe_trn_results), predictors(rfe_fit_results)))
print("Diffs predictors(rfe_fit_results) vs. predictors(rfe_trn_results):")
print(setdiff(predictors(rfe_fit_results), predictors(rfe_trn_results)))
}
}
if (grepl("Ensemble", glb_sel_mdl_id)) {
# Find which models are relevant
mdlimp_df <- subset(myget_feats_importance(glb_sel_mdl), importance > 5)
# Fit selected models on glb_trnobs_df
for (mdl_id in gsub(".prob", "",
gsub(glb_rsp_var_out, "", row.names(mdlimp_df), fixed = TRUE),
fixed = TRUE)) {
mdl_id_components <- unlist(strsplit(mdl_id, "[.]"))
mdlIdPfx <- paste0(c(head(mdl_id_components, -1), "Train"),
collapse = ".")
if (grepl("RFE\\.X\\.", mdlIdPfx))
mdlIndepVars <- myadjust_interaction_feats(myextract_actual_feats(
predictors(rfe_trn_results))) else
mdlIndepVars <- trim(unlist(
strsplit(glb_models_df[glb_models_df$id == mdl_id, "feats"], "[,]")))
ret_lst <-
myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = mdlIdPfx,
type = glb_model_type, tune.df = glb_tune_models_df,
trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds,
trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = tail(mdl_id_components, 1))),
indep_vars = mdlIndepVars,
rsp_var = glb_rsp_var,
fit_df = glb_trnobs_df, OOB_df = NULL)
glb_trnobs_df <- glb_get_predictions(df = glb_trnobs_df,
mdl_id = tail(glb_models_df$id, 1),
rsp_var_out = glb_rsp_var_out,
prob_threshold_def =
subset(glb_models_df, id == mdl_id)$opt.prob.threshold.OOB)
glb_newobs_df <- glb_get_predictions(df = glb_newobs_df,
mdl_id = tail(glb_models_df$id, 1),
rsp_var_out = glb_rsp_var_out,
prob_threshold_def =
subset(glb_models_df, id == mdl_id)$opt.prob.threshold.OOB)
}
}
print("***************")
print("Outliers in obsTrnDF not deleted yet")
print("***************")
# "Final" model
if ((model_method <- glb_sel_mdl$method) == "custom")
# get actual method from the mdl_id
model_method <- tail(unlist(strsplit(glb_sel_mdl_id, "[.]")), 1)
if (grepl("Ensemble", glb_sel_mdl_id)) {
# Find which models are relevant
mdlimp_df <- subset(myget_feats_importance(glb_sel_mdl), importance > 5)
if (glb_is_classification && glb_is_binomial)
indep_vars_vctr <- gsub("(.*)\\.(.*)\\.prob", "\\1\\.Train\\.\\2\\.prob",
row.names(mdlimp_df)) else
indep_vars_vctr <- gsub("(.*)\\.(.*)", "\\1\\.Train\\.\\2",
row.names(mdlimp_df))
} else indep_vars_vctr <-
trim(unlist(strsplit(glb_models_df[glb_models_df$id ==
glb_sel_mdl_id
, "feats"], "[,]")))
# Discontinuing use of tune_finmdl_df;
# since final model needs to be cved on glb_trnobs_df
# tune_finmdl_df <- NULL
# if (nrow(glb_sel_mdl$bestTune) > 0) {
# for (param in names(glb_sel_mdl$bestTune)) {
# #print(sprintf("param: %s", param))
# if (glb_sel_mdl$bestTune[1, param] != "none")
# tune_finmdl_df <- rbind(tune_finmdl_df,
# data.frame(parameter=param,
# min=glb_sel_mdl$bestTune[1, param],
# max=glb_sel_mdl$bestTune[1, param],
# by=1)) # by val does not matter
# }
# }
# Sync with parameters in mydsutils.R
#stop(here"); glb_to_sav(); glb_models_lst <- sav_models_lst; glb_models_df <- sav_models_df
if (!is.null(glb_preproc_methods) &&
((match_pos <- regexpr(gsub(".", "\\.",
paste(glb_preproc_methods, collapse = "|"),
fixed = TRUE), glb_sel_mdl_id)) != -1))
ths_preProcess <- str_sub(glb_sel_mdl_id, match_pos,
match_pos + attr(match_pos, "match.length") - 1) else
ths_preProcess <- NULL
mdl_id_pfx <- ifelse(grepl("Ensemble", glb_sel_mdl_id),
"Final.Ensemble", "Final")
trnobs_df <- if (is.null(glbObsTrnOutliers[[mdl_id_pfx]])) glb_trnobs_df else
glb_trnobs_df[!(glb_trnobs_df[, glb_id_var] %in%
glbObsTrnOutliers[[mdl_id_pfx]]), ]
# Force fitting of Final.glm to identify outliers
method_vctr <- unique(c("glm", tail(unlist(strsplit(glb_sel_mdl_id, "[.]")), 1)))
for (method in method_vctr) {
#source("caret_nominalTrainWorkflow.R")
# glmnet requires at least 2 indep vars
if ((length(indep_vars_vctr) == 1) && (method %in% "glmnet"))
next
ret_lst <-
myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst = list(
id.prefix = mdl_id_pfx,
type = glb_model_type, trainControl.method = "repeatedcv",
trainControl.number = glb_rcv_n_folds,
trainControl.repeats = glb_rcv_n_repeats,
trainControl.classProbs = glb_is_classification,
trainControl.summaryFunction = glbMdlMetricSummaryFn,
train.metric = glbMdlMetricSummary,
train.maximize = glbMdlMetricMaximize,
train.method = method,
train.preProcess = ths_preProcess)),
indep_vars = indep_vars_vctr, rsp_var = glb_rsp_var,
fit_df = trnobs_df, OOB_df = NULL)
}
if ((length(method_vctr) == 1) || (method != "glm")) {
glb_fin_mdl <- glb_models_lst[[length(glb_models_lst)]]
glb_fin_mdl_id <- glb_models_df[length(glb_models_lst), "id"]
}
}## [1] "***************"
## [1] "Outliers in obsTrnDF not deleted yet"
## [1] "***************"
## [1] "fitting model: Final.glm"
## [1] " indep_vars: biddable,sprice.d20nexp,spdiff.cut.fctr,D.ratio.wrds.stop.n.wrds.n,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,startprice.dcm1.is9,storage.fctr,D.chrs.pnct11.n.log,D.chrs.pnct13.n.log,startprice.dgt2.is9,color.fctr,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,prdl.my.fctr,D.chrs.uppr.n.log,D.chrs.n.log,D.wrds.n.log,D.terms.post.stem.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.log10,sprice.root2,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr"
## + Fold1.Rep1: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold1.Rep1: parameter=none
## + Fold2.Rep1: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold2.Rep1: parameter=none
## + Fold3.Rep1: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold3.Rep1: parameter=none
## + Fold1.Rep2: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold1.Rep2: parameter=none
## + Fold2.Rep2: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold2.Rep2: parameter=none
## + Fold3.Rep2: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold3.Rep2: parameter=none
## + Fold1.Rep3: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold1.Rep3: parameter=none
## + Fold2.Rep3: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold2.Rep3: parameter=none
## + Fold3.Rep3: parameter=none## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## - Fold3.Rep3: parameter=none
## Aggregating results
## Fitting final model on full training set
##
## Call:
## NULL
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.9681 -0.5779 -0.1363 0.4125 3.2961
##
## Coefficients: (26 not defined because of singularities)
## Estimate Std. Error z value
## (Intercept) 4.818e+00 1.980e+01 0.243
## .rnorm -2.241e-02 7.036e-02 -0.319
## D.chrs.n.log -1.851e+00 1.687e+00 -1.097
## D.chrs.pnct11.n.log -1.113e-01 2.756e-01 -0.404
## D.chrs.pnct13.n.log -2.204e-01 2.538e-01 -0.868
## D.chrs.uppr.n.log 2.824e-01 1.511e+00 0.187
## D.ratio.weight.sum.wrds.n -6.955e-03 2.420e-01 -0.029
## D.ratio.wrds.stop.n.wrds.n -6.235e+00 2.607e+00 -2.392
## D.terms.post.stem.n.log 1.221e+01 6.344e+00 1.925
## D.terms.post.stop.n.log -1.692e+01 6.386e+00 -2.649
## D.weight.post.stem.sum -1.434e+00 3.556e+00 -0.403
## D.weight.post.stop.sum 1.288e+00 3.448e+00 0.374
## D.weight.sum NA NA NA
## D.weight.sum.stem.stop.Ratio 2.281e+00 1.993e+01 0.114
## D.wrds.n.log 5.303e+00 1.614e+00 3.285
## D.wrds.stop.n.log 2.504e-01 4.553e-01 0.550
## D.wrds.unq.n.log NA NA NA
## biddable 2.218e+00 1.932e-01 11.478
## cellular.fctr1 1.632e-01 2.387e-01 0.684
## cellular.fctrUnknown -5.645e-01 3.629e-01 -1.555
## color.fctrGold -2.126e-01 4.745e-01 -0.448
## `color.fctrSpace Gray` -9.572e-03 2.867e-01 -0.033
## color.fctrUnknown -7.209e-02 1.980e-01 -0.364
## color.fctrWhite -1.971e-01 2.186e-01 -0.901
## `condition.fctrFor parts or not working` -6.153e-01 2.959e-01 -2.079
## `condition.fctrManufacturer refurbished` -2.487e-01 5.034e-01 -0.494
## condition.fctrNew 1.134e-01 2.671e-01 0.425
## `condition.fctrNew other (see details)` 6.755e-01 3.804e-01 1.776
## `condition.fctrSeller refurbished` -7.642e-01 3.297e-01 -2.318
## prdl.my.fctriPad1 -8.169e-01 4.361e-01 -1.873
## prdl.my.fctriPad2 -4.572e-01 3.849e-01 -1.188
## prdl.my.fctriPad3 1.912e-01 4.248e-01 0.450
## prdl.my.fctriPad4 8.521e-01 4.467e-01 1.908
## prdl.my.fctriPadAir 1.231e+00 4.419e-01 2.786
## prdl.my.fctriPadAir2 1.508e+00 4.973e-01 3.032
## prdl.my.fctriPadmini -3.018e-01 3.801e-01 -0.794
## prdl.my.fctriPadmini2 5.812e-01 4.593e-01 1.265
## prdl.my.fctriPadmini3 4.597e-01 5.288e-01 0.869
## `spdiff.cut.fctr(-100,-10]` 2.286e+00 2.845e-01 8.035
## `spdiff.cut.fctr(-10,-1]` 4.108e+00 3.673e-01 11.184
## `spdiff.cut.fctr(-1,0]` 5.699e+00 7.320e-01 7.786
## `spdiff.cut.fctr(0,1]` 4.867e+00 8.097e-01 6.012
## `spdiff.cut.fctr(1,10]` 3.920e+00 3.762e-01 10.419
## `spdiff.cut.fctr(10,100]` 3.253e+00 3.340e-01 9.739
## `spdiff.cut.fctr(100,1e+03]` 1.174e+00 5.228e-01 2.246
## sprice.d20nexp -1.820e+00 1.815e+00 -1.002
## sprice.log10 -5.453e-01 5.075e-01 -1.074
## sprice.root2 -8.693e-02 7.844e-02 -1.108
## startprice.dcm1.is9 -6.420e-01 2.678e-01 -2.397
## startprice.dcm2.is9 1.896e-01 2.660e-01 0.713
## startprice.dgt1.is9 7.679e-02 1.880e-01 0.409
## startprice.dgt2.is9 6.693e-02 2.192e-01 0.305
## storage.fctr16 -7.430e-01 4.403e-01 -1.687
## storage.fctr32 -7.296e-01 4.546e-01 -1.605
## storage.fctr64 -1.910e-01 4.343e-01 -0.440
## storage.fctrUnknown -9.796e-02 5.593e-01 -0.175
## `cellular.fctr0:carrier.fctrNone` NA NA NA
## `cellular.fctr1:carrier.fctrNone` NA NA NA
## `cellular.fctrUnknown:carrier.fctrNone` NA NA NA
## `cellular.fctr0:carrier.fctrOther` NA NA NA
## `cellular.fctr1:carrier.fctrOther` 1.364e+01 1.691e+03 0.008
## `cellular.fctrUnknown:carrier.fctrOther` NA NA NA
## `cellular.fctr0:carrier.fctrSprint` NA NA NA
## `cellular.fctr1:carrier.fctrSprint` 4.476e-01 5.582e-01 0.802
## `cellular.fctrUnknown:carrier.fctrSprint` NA NA NA
## `cellular.fctr0:carrier.fctrT-Mobile` NA NA NA
## `cellular.fctr1:carrier.fctrT-Mobile` 2.231e-01 6.784e-01 0.329
## `cellular.fctrUnknown:carrier.fctrT-Mobile` NA NA NA
## `cellular.fctr0:carrier.fctrUnknown` NA NA NA
## `cellular.fctr1:carrier.fctrUnknown` -3.460e-02 4.030e-01 -0.086
## `cellular.fctrUnknown:carrier.fctrUnknown` NA NA NA
## `cellular.fctr0:carrier.fctrVerizon` NA NA NA
## `cellular.fctr1:carrier.fctrVerizon` 1.447e-01 3.391e-01 0.427
## `cellular.fctrUnknown:carrier.fctrVerizon` NA NA NA
## `prdl.my.fctrUnknown:.clusterid.fctr2` 8.204e-01 6.142e-01 1.336
## `prdl.my.fctriPad1:.clusterid.fctr2` -1.065e+00 5.999e-01 -1.776
## `prdl.my.fctriPad2:.clusterid.fctr2` NA NA NA
## `prdl.my.fctriPad3:.clusterid.fctr2` -8.253e-01 7.431e-01 -1.111
## `prdl.my.fctriPad4:.clusterid.fctr2` -2.964e-01 7.953e-01 -0.373
## `prdl.my.fctriPadAir:.clusterid.fctr2` 4.084e-01 5.991e-01 0.682
## `prdl.my.fctriPadAir2:.clusterid.fctr2` NA NA NA
## `prdl.my.fctriPadmini:.clusterid.fctr2` NA NA NA
## `prdl.my.fctriPadmini2:.clusterid.fctr2` -6.078e-01 8.755e-01 -0.694
## `prdl.my.fctriPadmini3:.clusterid.fctr2` NA NA NA
## `prdl.my.fctrUnknown:.clusterid.fctr3` 6.544e-02 8.082e-01 0.081
## `prdl.my.fctriPad1:.clusterid.fctr3` -1.499e+00 7.064e-01 -2.121
## `prdl.my.fctriPad2:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPad3:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPad4:.clusterid.fctr3` -1.618e+01 3.993e+02 -0.041
## `prdl.my.fctriPadAir:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPadAir2:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPadmini:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPadmini2:.clusterid.fctr3` NA NA NA
## `prdl.my.fctriPadmini3:.clusterid.fctr3` NA NA NA
## Pr(>|z|)
## (Intercept) 0.80769
## .rnorm 0.75004
## D.chrs.n.log 0.27271
## D.chrs.pnct11.n.log 0.68629
## D.chrs.pnct13.n.log 0.38525
## D.chrs.uppr.n.log 0.85178
## D.ratio.weight.sum.wrds.n 0.97708
## D.ratio.wrds.stop.n.wrds.n 0.01676 *
## D.terms.post.stem.n.log 0.05428 .
## D.terms.post.stop.n.log 0.00807 **
## D.weight.post.stem.sum 0.68672
## D.weight.post.stop.sum 0.70870
## D.weight.sum NA
## D.weight.sum.stem.stop.Ratio 0.90889
## D.wrds.n.log 0.00102 **
## D.wrds.stop.n.log 0.58241
## D.wrds.unq.n.log NA
## biddable < 2e-16 ***
## cellular.fctr1 0.49420
## cellular.fctrUnknown 0.11987
## color.fctrGold 0.65405
## `color.fctrSpace Gray` 0.97336
## color.fctrUnknown 0.71579
## color.fctrWhite 0.36733
## `condition.fctrFor parts or not working` 0.03760 *
## `condition.fctrManufacturer refurbished` 0.62137
## condition.fctrNew 0.67100
## `condition.fctrNew other (see details)` 0.07575 .
## `condition.fctrSeller refurbished` 0.02045 *
## prdl.my.fctriPad1 0.06103 .
## prdl.my.fctriPad2 0.23493
## prdl.my.fctriPad3 0.65259
## prdl.my.fctriPad4 0.05642 .
## prdl.my.fctriPadAir 0.00533 **
## prdl.my.fctriPadAir2 0.00243 **
## prdl.my.fctriPadmini 0.42718
## prdl.my.fctriPadmini2 0.20571
## prdl.my.fctriPadmini3 0.38471
## `spdiff.cut.fctr(-100,-10]` 9.38e-16 ***
## `spdiff.cut.fctr(-10,-1]` < 2e-16 ***
## `spdiff.cut.fctr(-1,0]` 6.93e-15 ***
## `spdiff.cut.fctr(0,1]` 1.84e-09 ***
## `spdiff.cut.fctr(1,10]` < 2e-16 ***
## `spdiff.cut.fctr(10,100]` < 2e-16 ***
## `spdiff.cut.fctr(100,1e+03]` 0.02472 *
## sprice.d20nexp 0.31610
## sprice.log10 0.28263
## sprice.root2 0.26778
## startprice.dcm1.is9 0.01653 *
## startprice.dcm2.is9 0.47601
## startprice.dgt1.is9 0.68290
## startprice.dgt2.is9 0.76007
## storage.fctr16 0.09154 .
## storage.fctr32 0.10845
## storage.fctr64 0.66013
## storage.fctrUnknown 0.86096
## `cellular.fctr0:carrier.fctrNone` NA
## `cellular.fctr1:carrier.fctrNone` NA
## `cellular.fctrUnknown:carrier.fctrNone` NA
## `cellular.fctr0:carrier.fctrOther` NA
## `cellular.fctr1:carrier.fctrOther` 0.99357
## `cellular.fctrUnknown:carrier.fctrOther` NA
## `cellular.fctr0:carrier.fctrSprint` NA
## `cellular.fctr1:carrier.fctrSprint` 0.42261
## `cellular.fctrUnknown:carrier.fctrSprint` NA
## `cellular.fctr0:carrier.fctrT-Mobile` NA
## `cellular.fctr1:carrier.fctrT-Mobile` 0.74229
## `cellular.fctrUnknown:carrier.fctrT-Mobile` NA
## `cellular.fctr0:carrier.fctrUnknown` NA
## `cellular.fctr1:carrier.fctrUnknown` 0.93159
## `cellular.fctrUnknown:carrier.fctrUnknown` NA
## `cellular.fctr0:carrier.fctrVerizon` NA
## `cellular.fctr1:carrier.fctrVerizon` 0.66957
## `cellular.fctrUnknown:carrier.fctrVerizon` NA
## `prdl.my.fctrUnknown:.clusterid.fctr2` 0.18165
## `prdl.my.fctriPad1:.clusterid.fctr2` 0.07578 .
## `prdl.my.fctriPad2:.clusterid.fctr2` NA
## `prdl.my.fctriPad3:.clusterid.fctr2` 0.26669
## `prdl.my.fctriPad4:.clusterid.fctr2` 0.70943
## `prdl.my.fctriPadAir:.clusterid.fctr2` 0.49539
## `prdl.my.fctriPadAir2:.clusterid.fctr2` NA
## `prdl.my.fctriPadmini:.clusterid.fctr2` NA
## `prdl.my.fctriPadmini2:.clusterid.fctr2` 0.48751
## `prdl.my.fctriPadmini3:.clusterid.fctr2` NA
## `prdl.my.fctrUnknown:.clusterid.fctr3` 0.93547
## `prdl.my.fctriPad1:.clusterid.fctr3` 0.03389 *
## `prdl.my.fctriPad2:.clusterid.fctr3` NA
## `prdl.my.fctriPad3:.clusterid.fctr3` NA
## `prdl.my.fctriPad4:.clusterid.fctr3` 0.96768
## `prdl.my.fctriPadAir:.clusterid.fctr3` NA
## `prdl.my.fctriPadAir2:.clusterid.fctr3` NA
## `prdl.my.fctriPadmini:.clusterid.fctr3` NA
## `prdl.my.fctriPadmini2:.clusterid.fctr3` NA
## `prdl.my.fctriPadmini3:.clusterid.fctr3` NA
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 2554.0 on 1849 degrees of freedom
## Residual deviance: 1329.2 on 1782 degrees of freedom
## AIC: 1465.2
##
## Number of Fisher Scoring iterations: 15## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading## Warning in predict.lm(object, newdata, se.fit, scale = 1, type =
## ifelse(type == : prediction from a rank-deficient fit may be misleading
## sold.fctr sold.fctr.predict.Final.glm.N sold.fctr.predict.Final.glm.Y
## 1 N 833 162
## 2 Y 138 717
## Prediction
## Reference N Y
## N 833 162
## Y 138 717
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 8.378378e-01 6.744505e-01 8.202383e-01 8.543588e-01 5.378378e-01
## AccuracyPValue McnemarPValue
## 3.770097e-164 1.842093e-01
## id
## 1 Final.glm
## feats
## 1 biddable,sprice.d20nexp,spdiff.cut.fctr,D.ratio.wrds.stop.n.wrds.n,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,startprice.dcm1.is9,storage.fctr,D.chrs.pnct11.n.log,D.chrs.pnct13.n.log,startprice.dgt2.is9,color.fctr,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,prdl.my.fctr,D.chrs.uppr.n.log,D.chrs.n.log,D.wrds.n.log,D.terms.post.stem.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.log10,sprice.root2,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 1 4.081 0.326
## max.AUCpROC.fit max.Sens.fit max.Spec.fit max.AUCROCR.fit
## 1 0.8376238 0.8904523 0.7847953 0.9180828
## opt.prob.threshold.fit max.f.score.fit max.Accuracy.fit
## 1 0.4 0.8269896 0.8185584
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.8202383 0.8543588 0.6330313
## max.AccuracySD.fit max.KappaSD.fit
## 1 0.01101111 0.0227628
## [1] "fitting model: Final.glmnet"
## [1] " indep_vars: biddable,sprice.d20nexp,spdiff.cut.fctr,D.ratio.wrds.stop.n.wrds.n,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,startprice.dcm1.is9,storage.fctr,D.chrs.pnct11.n.log,D.chrs.pnct13.n.log,startprice.dgt2.is9,color.fctr,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,prdl.my.fctr,D.chrs.uppr.n.log,D.chrs.n.log,D.wrds.n.log,D.terms.post.stem.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.log10,sprice.root2,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr"
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 1, lambda = 0.00548 on full training set## Warning in myfit_mdl(mdl_specs_lst = myinit_mdl_specs_lst(mdl_specs_lst
## = list(id.prefix = mdl_id_pfx, : model's bestTune found at an extreme of
## tuneGrid for parameter: alpha
## Length Class Mode
## a0 85 -none- numeric
## beta 7905 dgCMatrix S4
## df 85 -none- numeric
## dim 2 -none- numeric
## lambda 85 -none- numeric
## dev.ratio 85 -none- numeric
## nulldev 1 -none- numeric
## npasses 1 -none- numeric
## jerr 1 -none- numeric
## offset 1 -none- logical
## classnames 2 -none- character
## call 5 -none- call
## nobs 1 -none- numeric
## lambdaOpt 1 -none- numeric
## xNames 93 -none- character
## problemType 1 -none- character
## tuneValue 2 data.frame list
## obsLevels 2 -none- character
## [1] "min lambda > lambdaOpt:"
## (Intercept)
## -0.5748424332
## D.chrs.pnct13.n.log
## -0.0154166095
## D.ratio.weight.sum.wrds.n
## -0.0004393607
## D.terms.post.stop.n.log
## -0.0089438359
## biddable
## 1.9394377386
## cellular.fctr1
## 0.0702567191
## cellular.fctrUnknown
## -0.1320135490
## color.fctrWhite
## -0.0266151806
## condition.fctrFor parts or not working
## -0.2200229713
## condition.fctrNew other (see details)
## 0.4387934720
## condition.fctrSeller refurbished
## -0.4201823354
## prdl.my.fctriPad1
## -0.4548555560
## prdl.my.fctriPad2
## -0.3245612053
## prdl.my.fctriPad4
## 0.1977145978
## prdl.my.fctriPadAir
## 0.6128927100
## prdl.my.fctriPadAir2
## 0.8968352444
## prdl.my.fctriPadmini
## -0.1622395269
## prdl.my.fctriPadmini2
## 0.1674866134
## spdiff.cut.fctr(-100,-10]
## 1.1267058623
## spdiff.cut.fctr(-10,-1]
## 2.6792229681
## spdiff.cut.fctr(-1,0]
## 3.6061648196
## spdiff.cut.fctr(0,1]
## 2.7445701555
## spdiff.cut.fctr(1,10]
## 2.4590410900
## spdiff.cut.fctr(10,100]
## 2.0111381095
## spdiff.cut.fctr(100,1e+03]
## 0.2412671988
## sprice.log10
## -0.1516075270
## sprice.root2
## -0.0864776262
## startprice.dcm1.is9
## -0.2865501177
## storage.fctr32
## -0.0032900963
## storage.fctr64
## 0.2388740478
## cellular.fctr1:carrier.fctrSprint
## 0.1795038991
## cellular.fctr1:carrier.fctrT-Mobile
## 0.0553779569
## cellular.fctrUnknown:carrier.fctrUnknown
## -0.0076892795
## prdl.my.fctrUnknown:.clusterid.fctr2
## 0.3891983721
## prdl.my.fctriPad1:.clusterid.fctr2
## -0.4118967518
## prdl.my.fctriPad3:.clusterid.fctr2
## -0.2066322688
## prdl.my.fctriPad1:.clusterid.fctr3
## -0.7279651218
## prdl.my.fctriPad4:.clusterid.fctr3
## -2.2439690983
## [1] "max lambda < lambdaOpt:"
## (Intercept)
## -0.4887481635
## D.chrs.pnct11.n.log
## -0.0049897588
## D.chrs.pnct13.n.log
## -0.0343205066
## D.ratio.weight.sum.wrds.n
## -0.0065169343
## D.terms.post.stop.n.log
## -0.0039264020
## D.weight.sum.stem.stop.Ratio
## -0.1489884401
## biddable
## 1.9631938033
## cellular.fctr1
## 0.0786061208
## cellular.fctrUnknown
## -0.1412878868
## color.fctrWhite
## -0.0385898938
## condition.fctrFor parts or not working
## -0.2344565867
## condition.fctrNew other (see details)
## 0.4627975233
## condition.fctrSeller refurbished
## -0.4454800069
## prdl.my.fctriPad1
## -0.4810735152
## prdl.my.fctriPad2
## -0.3463142796
## prdl.my.fctriPad4
## 0.2270111894
## prdl.my.fctriPadAir
## 0.6385121785
## prdl.my.fctriPadAir2
## 0.9150943539
## prdl.my.fctriPadmini
## -0.1801302700
## prdl.my.fctriPadmini2
## 0.1762468108
## spdiff.cut.fctr(-100,-10]
## 1.2080758350
## spdiff.cut.fctr(-10,-1]
## 2.7783245058
## spdiff.cut.fctr(-1,0]
## 3.7489075715
## spdiff.cut.fctr(0,1]
## 2.8862595318
## spdiff.cut.fctr(1,10]
## 2.5615257475
## spdiff.cut.fctr(10,100]
## 2.0983774318
## spdiff.cut.fctr(100,1e+03]
## 0.3228440594
## sprice.log10
## -0.1541621517
## sprice.root2
## -0.0866668571
## startprice.dcm1.is9
## -0.2961675551
## storage.fctr32
## -0.0092455272
## storage.fctr64
## 0.2513803695
## cellular.fctr1:carrier.fctrSprint
## 0.2045698195
## cellular.fctr1:carrier.fctrT-Mobile
## 0.0719134607
## cellular.fctrUnknown:carrier.fctrUnknown
## -0.0119069939
## prdl.my.fctrUnknown:.clusterid.fctr2
## 0.4302836448
## prdl.my.fctriPad1:.clusterid.fctr2
## -0.4315400934
## prdl.my.fctriPad3:.clusterid.fctr2
## -0.2494241814
## prdl.my.fctrUnknown:.clusterid.fctr3
## -0.0005985027
## prdl.my.fctriPad1:.clusterid.fctr3
## -0.7691702044
## prdl.my.fctriPad4:.clusterid.fctr3
## -2.4003871462
## sold.fctr sold.fctr.predict.Final.glmnet.N
## 1 N 831
## 2 Y 138
## sold.fctr.predict.Final.glmnet.Y
## 1 164
## 2 717
## Prediction
## Reference N Y
## N 831 164
## Y 138 717
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNull
## 8.367568e-01 6.723340e-01 8.191140e-01 8.533244e-01 5.378378e-01
## AccuracyPValue McnemarPValue
## 7.369741e-163 1.502672e-01
## id
## 1 Final.glmnet
## feats
## 1 biddable,sprice.d20nexp,spdiff.cut.fctr,D.ratio.wrds.stop.n.wrds.n,startprice.dcm2.is9,D.wrds.stop.n.log,D.weight.sum.stem.stop.Ratio,D.ratio.weight.sum.wrds.n,.rnorm,startprice.dcm1.is9,storage.fctr,D.chrs.pnct11.n.log,D.chrs.pnct13.n.log,startprice.dgt2.is9,color.fctr,D.weight.post.stop.sum,D.weight.post.stem.sum,D.weight.sum,prdl.my.fctr,D.chrs.uppr.n.log,D.chrs.n.log,D.wrds.n.log,D.terms.post.stem.n.log,D.wrds.unq.n.log,D.terms.post.stop.n.log,cellular.fctr,startprice.dgt1.is9,condition.fctr,sprice.log10,sprice.root2,cellular.fctr:carrier.fctr,prdl.my.fctr:.clusterid.fctr
## max.nTuningRuns min.elapsedtime.everything min.elapsedtime.final
## 1 25 17.71 1.915
## max.AUCpROC.fit max.Sens.fit max.Spec.fit max.AUCROCR.fit
## 1 0.8279468 0.9015075 0.754386 0.908196
## opt.prob.threshold.fit max.f.score.fit max.Accuracy.fit
## 1 0.4 0.8260369 0.8203656
## max.AccuracyLower.fit max.AccuracyUpper.fit max.Kappa.fit
## 1 0.819114 0.8533244 0.6357114
## max.AccuracySD.fit max.KappaSD.fit
## 1 0.01119409 0.02374578rm(ret_lst)
glb_chunks_df <- myadd_chunk(glb_chunks_df, "fit.data.training", major.inc=FALSE)## label step_major step_minor label_minor bgn end
## 14 fit.data.training 7 0 0 504.742 533.21
## 15 fit.data.training 7 1 1 533.211 NA
## elapsed
## 14 28.468
## 15 NA#stop(here"); glb_to_sav()
if (glb_is_classification && glb_is_binomial)
prob_threshold <- glb_models_df[glb_models_df$id == glb_sel_mdl_id,
"opt.prob.threshold.OOB"] else
prob_threshold <- NULL
if (grepl("Ensemble", glb_fin_mdl_id)) {
# Get predictions for each model in ensemble; Outliers that have been moved to OOB might not have been predicted yet
mdlEnsembleComps <- unlist(str_split(subset(glb_models_df,
id == glb_fin_mdl_id)$feats, ","))
if (glb_is_classification && glb_is_binomial)
mdlEnsembleComps <- gsub("\\.prob$", "", mdlEnsembleComps)
mdlEnsembleComps <- gsub(paste0("^",
gsub(".", "\\.", glb_rsp_var_out, fixed = TRUE)),
"", mdlEnsembleComps)
for (mdl_id in mdlEnsembleComps) {
glb_trnobs_df <- glb_get_predictions(df=glb_trnobs_df, mdl_id=mdl_id,
rsp_var_out=glb_rsp_var_out,
prob_threshold_def=prob_threshold)
glb_newobs_df <- glb_get_predictions(df=glb_newobs_df, mdl_id=mdl_id,
rsp_var_out=glb_rsp_var_out,
prob_threshold_def=prob_threshold)
}
}
glb_trnobs_df <- glb_get_predictions(df=glb_trnobs_df, mdl_id=glb_fin_mdl_id,
rsp_var_out=glb_rsp_var_out,
prob_threshold_def=prob_threshold)## Warning in glb_get_predictions(df = glb_trnobs_df, mdl_id =
## glb_fin_mdl_id, : Using default probability threshold: 0.3glb_featsimp_df <- myget_feats_importance(mdl=glb_fin_mdl,
featsimp_df=glb_featsimp_df)
glb_featsimp_df[, paste0(glb_fin_mdl_id, ".importance")] <- glb_featsimp_df$importance
print(glb_featsimp_df)## All.X.glmnet.importance
## spdiff.cut.fctr(-1,0] 100.00000
## spdiff.cut.fctr(0,1] 91.77769
## spdiff.cut.fctr(-10,-1] 89.53891
## spdiff.cut.fctr(1,10] 91.77825
## spdiff.cut.fctr(10,100] 86.86797
## biddable 75.17563
## spdiff.cut.fctr(-100,-10] 73.61063
## prdl.my.fctriPadAir2 69.61045
## prdl.my.fctriPadAir 66.43833
## condition.fctrNew other (see details) 59.12005
## prdl.my.fctrUnknown:.clusterid.fctr2 60.62908
## spdiff.cut.fctr(100,1e+03] 62.38931
## storage.fctr64 54.64454
## prdl.my.fctriPad4 65.16814
## cellular.fctr1:carrier.fctrSprint 61.81532
## prdl.my.fctriPadmini2 64.52272
## cellular.fctr1 52.71142
## cellular.fctr1:carrier.fctrT-Mobile 50.21020
## .rnorm 52.38568
## D.chrs.n.log 50.18255
## D.chrs.uppr.n.log 52.60858
## D.ratio.wrds.stop.n.wrds.n 26.94246
## D.terms.post.stem.n.log 50.17209
## D.weight.post.stem.sum 51.63919
## D.weight.post.stop.sum 52.71142
## D.weight.sum 51.66462
## D.wrds.n.log 63.88632
## D.wrds.stop.n.log 53.73265
## D.wrds.unq.n.log 51.14189
## cellular.fctr0:carrier.fctrNone 54.22434
## cellular.fctr0:carrier.fctrOther 52.71142
## cellular.fctr0:carrier.fctrSprint 52.71142
## cellular.fctr0:carrier.fctrT-Mobile 52.71142
## cellular.fctr0:carrier.fctrUnknown 52.71142
## cellular.fctr0:carrier.fctrVerizon 52.71142
## cellular.fctr1:carrier.fctrNone 52.71142
## cellular.fctr1:carrier.fctrOther 75.82223
## cellular.fctr1:carrier.fctrUnknown 52.55260
## cellular.fctr1:carrier.fctrVerizon 55.50911
## cellular.fctrUnknown:carrier.fctrNone 52.71142
## cellular.fctrUnknown:carrier.fctrOther 52.71142
## cellular.fctrUnknown:carrier.fctrSprint 52.71142
## cellular.fctrUnknown:carrier.fctrT-Mobile 52.71142
## cellular.fctrUnknown:carrier.fctrVerizon 52.71142
## color.fctrGold 53.48381
## color.fctrSpace Gray 49.63204
## color.fctrUnknown 55.86778
## condition.fctrManufacturer refurbished 42.10131
## condition.fctrNew 52.97506
## prdl.my.fctriPad2:.clusterid.fctr2 52.71142
## prdl.my.fctriPad2:.clusterid.fctr3 52.71142
## prdl.my.fctriPad3 54.34045
## prdl.my.fctriPad3:.clusterid.fctr3 52.71142
## prdl.my.fctriPad4:.clusterid.fctr2 39.81232
## prdl.my.fctriPadAir2:.clusterid.fctr2 52.71142
## prdl.my.fctriPadAir2:.clusterid.fctr3 52.71142
## prdl.my.fctriPadAir:.clusterid.fctr2 49.80426
## prdl.my.fctriPadAir:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini2:.clusterid.fctr2 47.13793
## prdl.my.fctriPadmini2:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini3 59.42417
## prdl.my.fctriPadmini3:.clusterid.fctr2 52.71142
## prdl.my.fctriPadmini3:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini:.clusterid.fctr2 52.71142
## prdl.my.fctriPadmini:.clusterid.fctr3 52.71142
## sprice.d20nexp 38.83227
## startprice.dcm2.is9 51.96783
## startprice.dgt1.is9 52.93673
## startprice.dgt2.is9 53.73722
## storage.fctr16 51.55882
## storage.fctrUnknown 68.07849
## prdl.my.fctrUnknown:.clusterid.fctr3 45.97618
## D.chrs.pnct11.n.log 56.89033
## D.ratio.weight.sum.wrds.n 52.18561
## storage.fctr32 52.71638
## D.weight.sum.stem.stop.Ratio 26.99368
## cellular.fctrUnknown:carrier.fctrUnknown 45.72360
## D.terms.post.stop.n.log 44.49862
## D.chrs.pnct13.n.log 51.24550
## color.fctrWhite 50.71378
## sprice.root2 51.74950
## cellular.fctrUnknown 45.68888
## sprice.log10 47.59704
## prdl.my.fctriPadmini 48.49638
## prdl.my.fctriPad3:.clusterid.fctr2 55.10589
## condition.fctrFor parts or not working 46.18220
## startprice.dcm1.is9 48.37251
## prdl.my.fctriPad2 48.61429
## prdl.my.fctriPad1:.clusterid.fctr2 46.14804
## condition.fctrSeller refurbished 43.60662
## prdl.my.fctriPad1 46.63375
## prdl.my.fctriPad1:.clusterid.fctr3 42.66158
## prdl.my.fctriPad4:.clusterid.fctr3 0.00000
## importance
## spdiff.cut.fctr(-1,0] 100.00000
## spdiff.cut.fctr(0,1] 85.31015
## spdiff.cut.fctr(-10,-1] 84.15852
## spdiff.cut.fctr(1,10] 80.40773
## spdiff.cut.fctr(10,100] 72.75819
## biddable 71.47982
## spdiff.cut.fctr(-100,-10] 57.67475
## prdl.my.fctriPadAir2 53.70014
## prdl.my.fctriPadAir 48.86583
## condition.fctrNew other (see details) 45.89626
## prdl.my.fctrUnknown:.clusterid.fctr2 45.06580
## spdiff.cut.fctr(100,1e+03] 42.57950
## storage.fctr64 42.47779
## prdl.my.fctriPad4 41.79089
## cellular.fctr1:carrier.fctrSprint 41.47670
## prdl.my.fctriPadmini2 41.25744
## cellular.fctr1 39.59945
## cellular.fctr1:carrier.fctrT-Mobile 39.35301
## .rnorm 38.39431
## D.chrs.n.log 38.39431
## D.chrs.uppr.n.log 38.39431
## D.ratio.wrds.stop.n.wrds.n 38.39431
## D.terms.post.stem.n.log 38.39431
## D.weight.post.stem.sum 38.39431
## D.weight.post.stop.sum 38.39431
## D.weight.sum 38.39431
## D.wrds.n.log 38.39431
## D.wrds.stop.n.log 38.39431
## D.wrds.unq.n.log 38.39431
## cellular.fctr0:carrier.fctrNone 38.39431
## cellular.fctr0:carrier.fctrOther 38.39431
## cellular.fctr0:carrier.fctrSprint 38.39431
## cellular.fctr0:carrier.fctrT-Mobile 38.39431
## cellular.fctr0:carrier.fctrUnknown 38.39431
## cellular.fctr0:carrier.fctrVerizon 38.39431
## cellular.fctr1:carrier.fctrNone 38.39431
## cellular.fctr1:carrier.fctrOther 38.39431
## cellular.fctr1:carrier.fctrUnknown 38.39431
## cellular.fctr1:carrier.fctrVerizon 38.39431
## cellular.fctrUnknown:carrier.fctrNone 38.39431
## cellular.fctrUnknown:carrier.fctrOther 38.39431
## cellular.fctrUnknown:carrier.fctrSprint 38.39431
## cellular.fctrUnknown:carrier.fctrT-Mobile 38.39431
## cellular.fctrUnknown:carrier.fctrVerizon 38.39431
## color.fctrGold 38.39431
## color.fctrSpace Gray 38.39431
## color.fctrUnknown 38.39431
## condition.fctrManufacturer refurbished 38.39431
## condition.fctrNew 38.39431
## prdl.my.fctriPad2:.clusterid.fctr2 38.39431
## prdl.my.fctriPad2:.clusterid.fctr3 38.39431
## prdl.my.fctriPad3 38.39431
## prdl.my.fctriPad3:.clusterid.fctr3 38.39431
## prdl.my.fctriPad4:.clusterid.fctr2 38.39431
## prdl.my.fctriPadAir2:.clusterid.fctr2 38.39431
## prdl.my.fctriPadAir2:.clusterid.fctr3 38.39431
## prdl.my.fctriPadAir:.clusterid.fctr2 38.39431
## prdl.my.fctriPadAir:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini2:.clusterid.fctr2 38.39431
## prdl.my.fctriPadmini2:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini3 38.39431
## prdl.my.fctriPadmini3:.clusterid.fctr2 38.39431
## prdl.my.fctriPadmini3:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini:.clusterid.fctr2 38.39431
## prdl.my.fctriPadmini:.clusterid.fctr3 38.39431
## sprice.d20nexp 38.39431
## startprice.dcm2.is9 38.39431
## startprice.dgt1.is9 38.39431
## startprice.dgt2.is9 38.39431
## storage.fctr16 38.39431
## storage.fctrUnknown 38.39431
## prdl.my.fctrUnknown:.clusterid.fctr3 38.39379
## D.chrs.pnct11.n.log 38.38991
## D.ratio.weight.sum.wrds.n 38.38146
## storage.fctr32 38.33297
## D.weight.sum.stem.stop.Ratio 38.26292
## cellular.fctrUnknown:carrier.fctrUnknown 38.25950
## D.terms.post.stop.n.log 38.24626
## D.chrs.pnct13.n.log 38.11481
## color.fctrWhite 37.93000
## sprice.root2 36.91983
## cellular.fctrUnknown 36.13550
## sprice.log10 35.80737
## prdl.my.fctriPadmini 35.61259
## prdl.my.fctriPad3:.clusterid.fctr2 34.83380
## condition.fctrFor parts or not working 34.63052
## startprice.dcm1.is9 33.50057
## prdl.my.fctriPad2 32.84184
## prdl.my.fctriPad1:.clusterid.fctr2 31.35476
## condition.fctrSeller refurbished 31.20851
## prdl.my.fctriPad1 30.61657
## prdl.my.fctriPad1:.clusterid.fctr3 25.94724
## prdl.my.fctriPad4:.clusterid.fctr3 0.00000
## Final.glmnet.importance
## spdiff.cut.fctr(-1,0] 100.00000
## spdiff.cut.fctr(0,1] 85.31015
## spdiff.cut.fctr(-10,-1] 84.15852
## spdiff.cut.fctr(1,10] 80.40773
## spdiff.cut.fctr(10,100] 72.75819
## biddable 71.47982
## spdiff.cut.fctr(-100,-10] 57.67475
## prdl.my.fctriPadAir2 53.70014
## prdl.my.fctriPadAir 48.86583
## condition.fctrNew other (see details) 45.89626
## prdl.my.fctrUnknown:.clusterid.fctr2 45.06580
## spdiff.cut.fctr(100,1e+03] 42.57950
## storage.fctr64 42.47779
## prdl.my.fctriPad4 41.79089
## cellular.fctr1:carrier.fctrSprint 41.47670
## prdl.my.fctriPadmini2 41.25744
## cellular.fctr1 39.59945
## cellular.fctr1:carrier.fctrT-Mobile 39.35301
## .rnorm 38.39431
## D.chrs.n.log 38.39431
## D.chrs.uppr.n.log 38.39431
## D.ratio.wrds.stop.n.wrds.n 38.39431
## D.terms.post.stem.n.log 38.39431
## D.weight.post.stem.sum 38.39431
## D.weight.post.stop.sum 38.39431
## D.weight.sum 38.39431
## D.wrds.n.log 38.39431
## D.wrds.stop.n.log 38.39431
## D.wrds.unq.n.log 38.39431
## cellular.fctr0:carrier.fctrNone 38.39431
## cellular.fctr0:carrier.fctrOther 38.39431
## cellular.fctr0:carrier.fctrSprint 38.39431
## cellular.fctr0:carrier.fctrT-Mobile 38.39431
## cellular.fctr0:carrier.fctrUnknown 38.39431
## cellular.fctr0:carrier.fctrVerizon 38.39431
## cellular.fctr1:carrier.fctrNone 38.39431
## cellular.fctr1:carrier.fctrOther 38.39431
## cellular.fctr1:carrier.fctrUnknown 38.39431
## cellular.fctr1:carrier.fctrVerizon 38.39431
## cellular.fctrUnknown:carrier.fctrNone 38.39431
## cellular.fctrUnknown:carrier.fctrOther 38.39431
## cellular.fctrUnknown:carrier.fctrSprint 38.39431
## cellular.fctrUnknown:carrier.fctrT-Mobile 38.39431
## cellular.fctrUnknown:carrier.fctrVerizon 38.39431
## color.fctrGold 38.39431
## color.fctrSpace Gray 38.39431
## color.fctrUnknown 38.39431
## condition.fctrManufacturer refurbished 38.39431
## condition.fctrNew 38.39431
## prdl.my.fctriPad2:.clusterid.fctr2 38.39431
## prdl.my.fctriPad2:.clusterid.fctr3 38.39431
## prdl.my.fctriPad3 38.39431
## prdl.my.fctriPad3:.clusterid.fctr3 38.39431
## prdl.my.fctriPad4:.clusterid.fctr2 38.39431
## prdl.my.fctriPadAir2:.clusterid.fctr2 38.39431
## prdl.my.fctriPadAir2:.clusterid.fctr3 38.39431
## prdl.my.fctriPadAir:.clusterid.fctr2 38.39431
## prdl.my.fctriPadAir:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini2:.clusterid.fctr2 38.39431
## prdl.my.fctriPadmini2:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini3 38.39431
## prdl.my.fctriPadmini3:.clusterid.fctr2 38.39431
## prdl.my.fctriPadmini3:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini:.clusterid.fctr2 38.39431
## prdl.my.fctriPadmini:.clusterid.fctr3 38.39431
## sprice.d20nexp 38.39431
## startprice.dcm2.is9 38.39431
## startprice.dgt1.is9 38.39431
## startprice.dgt2.is9 38.39431
## storage.fctr16 38.39431
## storage.fctrUnknown 38.39431
## prdl.my.fctrUnknown:.clusterid.fctr3 38.39379
## D.chrs.pnct11.n.log 38.38991
## D.ratio.weight.sum.wrds.n 38.38146
## storage.fctr32 38.33297
## D.weight.sum.stem.stop.Ratio 38.26292
## cellular.fctrUnknown:carrier.fctrUnknown 38.25950
## D.terms.post.stop.n.log 38.24626
## D.chrs.pnct13.n.log 38.11481
## color.fctrWhite 37.93000
## sprice.root2 36.91983
## cellular.fctrUnknown 36.13550
## sprice.log10 35.80737
## prdl.my.fctriPadmini 35.61259
## prdl.my.fctriPad3:.clusterid.fctr2 34.83380
## condition.fctrFor parts or not working 34.63052
## startprice.dcm1.is9 33.50057
## prdl.my.fctriPad2 32.84184
## prdl.my.fctriPad1:.clusterid.fctr2 31.35476
## condition.fctrSeller refurbished 31.20851
## prdl.my.fctriPad1 30.61657
## prdl.my.fctriPad1:.clusterid.fctr3 25.94724
## prdl.my.fctriPad4:.clusterid.fctr3 0.00000if (glb_is_classification && glb_is_binomial)
glb_analytics_diag_plots(obs_df=glb_trnobs_df, mdl_id=glb_fin_mdl_id,
prob_threshold=glb_models_df[glb_models_df$id == glb_sel_mdl_id,
"opt.prob.threshold.OOB"]) else
glb_analytics_diag_plots(obs_df=glb_trnobs_df, mdl_id=glb_fin_mdl_id) ## Warning in glb_analytics_diag_plots(obs_df = glb_trnobs_df, mdl_id =
## glb_fin_mdl_id, : Limiting important feature scatter plots to 5 out of 32
## [1] "Min/Max Boundaries: "
## UniqueID sold.fctr sold.fctr.predict.All.X.glmnet.prob
## 15 10015 Y 0.2387295
## 27 10027 Y NA
## 2 10002 Y 0.9976313
## 23 10023 N NA
## 195 10195 N 0.1991452
## 1 10001 N 0.2771651
## 22 10022 N 0.5305515
## 20 10020 N 0.5492854
## 135 10135 N 0.6032367
## 127 10127 N NA
## 120 10120 N 0.7037805
## 182 10182 N 0.6858467
## 103 10103 N NA
## sold.fctr.predict.All.X.glmnet sold.fctr.predict.All.X.glmnet.err
## 15 N TRUE
## 27 <NA> NA
## 2 Y FALSE
## 23 <NA> NA
## 195 N FALSE
## 1 N FALSE
## 22 Y TRUE
## 20 Y TRUE
## 135 Y TRUE
## 127 <NA> NA
## 120 Y TRUE
## 182 Y TRUE
## 103 <NA> NA
## sold.fctr.predict.All.X.glmnet.err.abs
## 15 0.761270521
## 27 NA
## 2 0.002368685
## 23 NA
## 195 0.199145222
## 1 0.277165051
## 22 0.530551453
## 20 0.549285373
## 135 0.603236741
## 127 NA
## 120 0.703780514
## 182 0.685846730
## 103 NA
## sold.fctr.predict.All.X.glmnet.accurate
## 15 FALSE
## 27 NA
## 2 TRUE
## 23 NA
## 195 TRUE
## 1 TRUE
## 22 FALSE
## 20 FALSE
## 135 FALSE
## 127 NA
## 120 FALSE
## 182 FALSE
## 103 NA
## sold.fctr.predict.Final.glmnet.prob sold.fctr.predict.Final.glmnet
## 15 0.2374065 N
## 27 0.2891283 N
## 2 0.9867924 Y
## 23 0.3006404 Y
## 195 0.3549341 Y
## 1 0.4037304 Y
## 22 0.5224896 Y
## 20 0.5280866 Y
## 135 0.5554621 Y
## 127 0.6162385 Y
## 120 0.6390285 Y
## 182 0.6865033 Y
## 103 0.9114699 Y
## sold.fctr.predict.Final.glmnet.err
## 15 TRUE
## 27 TRUE
## 2 FALSE
## 23 TRUE
## 195 TRUE
## 1 TRUE
## 22 TRUE
## 20 TRUE
## 135 TRUE
## 127 TRUE
## 120 TRUE
## 182 TRUE
## 103 TRUE
## sold.fctr.predict.Final.glmnet.err.abs
## 15 0.76259347
## 27 0.71087169
## 2 0.01320757
## 23 0.30064037
## 195 0.35493408
## 1 0.40373038
## 22 0.52248959
## 20 0.52808658
## 135 0.55546206
## 127 0.61623850
## 120 0.63902847
## 182 0.68650334
## 103 0.91146990
## sold.fctr.predict.Final.glmnet.accurate
## 15 FALSE
## 27 FALSE
## 2 TRUE
## 23 FALSE
## 195 FALSE
## 1 FALSE
## 22 FALSE
## 20 FALSE
## 135 FALSE
## 127 FALSE
## 120 FALSE
## 182 FALSE
## 103 FALSE
## sold.fctr.predict.Final.glmnet.error .label
## 15 -0.062593475 10015
## 27 -0.010871691 10027
## 2 0.000000000 10002
## 23 0.000640367 10023
## 195 0.054934084 10195
## 1 0.103730378 10001
## 22 0.222489585 10022
## 20 0.228086584 10020
## 135 0.255462063 10135
## 127 0.316238503 10127
## 120 0.339028470 10120
## 182 0.386503340 10182
## 103 0.611469896 10103
## [1] "Inaccurate: "
## UniqueID sold.fctr sold.fctr.predict.All.X.glmnet.prob
## 1807 11817 Y NA
## 1688 11697 Y 0.003639011
## 1696 11705 Y NA
## 1352 11359 Y 0.038703606
## 332 10332 Y NA
## 1113 11119 Y 0.090092153
## sold.fctr.predict.All.X.glmnet sold.fctr.predict.All.X.glmnet.err
## 1807 <NA> NA
## 1688 N TRUE
## 1696 <NA> NA
## 1352 N TRUE
## 332 <NA> NA
## 1113 N TRUE
## sold.fctr.predict.All.X.glmnet.err.abs
## 1807 NA
## 1688 0.9963610
## 1696 NA
## 1352 0.9612964
## 332 NA
## 1113 0.9099078
## sold.fctr.predict.All.X.glmnet.accurate
## 1807 NA
## 1688 FALSE
## 1696 NA
## 1352 FALSE
## 332 NA
## 1113 FALSE
## sold.fctr.predict.Final.glmnet.prob sold.fctr.predict.Final.glmnet
## 1807 0.03420561 N
## 1688 0.03456082 N
## 1696 0.06394171 N
## 1352 0.06497112 N
## 332 0.06545247 N
## 1113 0.07378931 N
## sold.fctr.predict.Final.glmnet.err
## 1807 TRUE
## 1688 TRUE
## 1696 TRUE
## 1352 TRUE
## 332 TRUE
## 1113 TRUE
## sold.fctr.predict.Final.glmnet.err.abs
## 1807 0.9657944
## 1688 0.9654392
## 1696 0.9360583
## 1352 0.9350289
## 332 0.9345475
## 1113 0.9262107
## sold.fctr.predict.Final.glmnet.accurate
## 1807 FALSE
## 1688 FALSE
## 1696 FALSE
## 1352 FALSE
## 332 FALSE
## 1113 FALSE
## sold.fctr.predict.Final.glmnet.error
## 1807 -0.2657944
## 1688 -0.2654392
## 1696 -0.2360583
## 1352 -0.2350289
## 332 -0.2345475
## 1113 -0.2262107
## UniqueID sold.fctr sold.fctr.predict.All.X.glmnet.prob
## 1430 11437 N NA
## 925 10928 N NA
## 787 10788 N 0.5234147
## 1043 11046 N 0.3904027
## 1586 11595 N 0.4562730
## 1665 11674 N NA
## sold.fctr.predict.All.X.glmnet sold.fctr.predict.All.X.glmnet.err
## 1430 <NA> NA
## 925 <NA> NA
## 787 Y TRUE
## 1043 Y TRUE
## 1586 Y TRUE
## 1665 <NA> NA
## sold.fctr.predict.All.X.glmnet.err.abs
## 1430 NA
## 925 NA
## 787 0.5234147
## 1043 0.3904027
## 1586 0.4562730
## 1665 NA
## sold.fctr.predict.All.X.glmnet.accurate
## 1430 NA
## 925 NA
## 787 FALSE
## 1043 FALSE
## 1586 FALSE
## 1665 NA
## sold.fctr.predict.Final.glmnet.prob sold.fctr.predict.Final.glmnet
## 1430 0.3260384 Y
## 925 0.3293791 Y
## 787 0.3559182 Y
## 1043 0.3730374 Y
## 1586 0.3798302 Y
## 1665 0.9076423 Y
## sold.fctr.predict.Final.glmnet.err
## 1430 TRUE
## 925 TRUE
## 787 TRUE
## 1043 TRUE
## 1586 TRUE
## 1665 TRUE
## sold.fctr.predict.Final.glmnet.err.abs
## 1430 0.3260384
## 925 0.3293791
## 787 0.3559182
## 1043 0.3730374
## 1586 0.3798302
## 1665 0.9076423
## sold.fctr.predict.Final.glmnet.accurate
## 1430 FALSE
## 925 FALSE
## 787 FALSE
## 1043 FALSE
## 1586 FALSE
## 1665 FALSE
## sold.fctr.predict.Final.glmnet.error
## 1430 0.02603841
## 925 0.02937913
## 787 0.05591819
## 1043 0.07303739
## 1586 0.07983023
## 1665 0.60764233
## UniqueID sold.fctr sold.fctr.predict.All.X.glmnet.prob
## 1703 11712 N NA
## 1384 11391 N NA
## 490 10491 N NA
## 1760 11769 N 0.9523485
## 1682 11691 N 0.9822811
## 1464 11471 N NA
## sold.fctr.predict.All.X.glmnet sold.fctr.predict.All.X.glmnet.err
## 1703 <NA> NA
## 1384 <NA> NA
## 490 <NA> NA
## 1760 Y TRUE
## 1682 Y TRUE
## 1464 <NA> NA
## sold.fctr.predict.All.X.glmnet.err.abs
## 1703 NA
## 1384 NA
## 490 NA
## 1760 0.9523485
## 1682 0.9822811
## 1464 NA
## sold.fctr.predict.All.X.glmnet.accurate
## 1703 NA
## 1384 NA
## 490 NA
## 1760 FALSE
## 1682 FALSE
## 1464 NA
## sold.fctr.predict.Final.glmnet.prob sold.fctr.predict.Final.glmnet
## 1703 0.9265818 Y
## 1384 0.9305841 Y
## 490 0.9403422 Y
## 1760 0.9467505 Y
## 1682 0.9649070 Y
## 1464 0.9689273 Y
## sold.fctr.predict.Final.glmnet.err
## 1703 TRUE
## 1384 TRUE
## 490 TRUE
## 1760 TRUE
## 1682 TRUE
## 1464 TRUE
## sold.fctr.predict.Final.glmnet.err.abs
## 1703 0.9265818
## 1384 0.9305841
## 490 0.9403422
## 1760 0.9467505
## 1682 0.9649070
## 1464 0.9689273
## sold.fctr.predict.Final.glmnet.accurate
## 1703 FALSE
## 1384 FALSE
## 490 FALSE
## 1760 FALSE
## 1682 FALSE
## 1464 FALSE
## sold.fctr.predict.Final.glmnet.error
## 1703 0.6265818
## 1384 0.6305841
## 490 0.6403422
## 1760 0.6467505
## 1682 0.6649070
## 1464 0.6689273
dsp_feats_vctr <- c(NULL)
for(var in grep(".importance", names(glb_feats_df), fixed=TRUE, value=TRUE))
dsp_feats_vctr <- union(dsp_feats_vctr,
glb_feats_df[!is.na(glb_feats_df[, var]), "id"])
# print(glb_trnobs_df[glb_trnobs_df$UniqueID %in% FN_OOB_ids,
# grep(glb_rsp_var, names(glb_trnobs_df), value=TRUE)])
print(setdiff(names(glb_trnobs_df), names(glb_allobs_df)))## [1] "sold.fctr.predict.Final.glmnet.prob"
## [2] "sold.fctr.predict.Final.glmnet"
## [3] "sold.fctr.predict.Final.glmnet.err"
## [4] "sold.fctr.predict.Final.glmnet.err.abs"
## [5] "sold.fctr.predict.Final.glmnet.accurate"for (col in setdiff(names(glb_trnobs_df), names(glb_allobs_df)))
# Merge or cbind ?
glb_allobs_df[glb_allobs_df$.src == "Train", col] <- glb_trnobs_df[, col]
print(setdiff(names(glb_fitobs_df), names(glb_allobs_df)))## character(0)print(setdiff(names(glb_OOBobs_df), names(glb_allobs_df)))## character(0)for (col in setdiff(names(glb_OOBobs_df), names(glb_allobs_df)))
# Merge or cbind ?
glb_allobs_df[glb_allobs_df$.lcn == "OOB", col] <- glb_OOBobs_df[, col]
print(setdiff(names(glb_newobs_df), names(glb_allobs_df)))## character(0)if (glb_save_envir)
save(glb_feats_df, glb_allobs_df,
#glb_trnobs_df, glb_fitobs_df, glb_OOBobs_df, glb_newobs_df,
glb_models_df, dsp_models_df, glb_models_lst, glb_model_type,
glb_sel_mdl, glb_sel_mdl_id,
glb_fin_mdl, glb_fin_mdl_id,
file=paste0(glb_out_pfx, "dsk.RData"))
replay.petrisim(pn=glb_analytics_pn,
replay.trans=(glb_analytics_avl_objs <- c(glb_analytics_avl_objs,
"data.training.all.prediction","model.final")), flip_coord=TRUE)## time trans "bgn " "fit.data.training.all " "predict.data.new " "end "
## 0.0000 multiple enabled transitions: data.training.all data.new model.selected firing: data.training.all
## 1.0000 1 2 1 0 0
## 1.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction firing: data.new
## 2.0000 2 1 1 1 0
## 2.0000 multiple enabled transitions: data.training.all data.new model.selected model.final data.training.all.prediction data.new.prediction firing: model.selected
## 3.0000 3 0 2 1 0
## 3.0000 multiple enabled transitions: model.final data.training.all.prediction data.new.prediction firing: data.training.all.prediction
## 4.0000 5 0 1 1 1
## 4.0000 multiple enabled transitions: model.final data.training.all.prediction data.new.prediction firing: model.final
## 5.0000 4 0 0 2 1
glb_chunks_df <- myadd_chunk(glb_chunks_df, "predict.data.new", major.inc=TRUE)## label step_major step_minor label_minor bgn end
## 15 fit.data.training 7 1 1 533.211 541.249
## 16 predict.data.new 8 0 0 541.250 NA
## elapsed
## 15 8.038
## 16 NAStep 8.0: predict data new
# Compute final model predictions
#glb_to_sav(); all.equal(sav_allobs_df, glb_allobs_df); all.equal(sav_trnobs_df, glb_trnobs_df); all.equal(sav_newobs_df, glb_newobs_df)
if (glb_is_classification && glb_is_binomial)
prob_threshold_def <-
glb_models_df[glb_models_df$id == glb_sel_mdl_id, "opt.prob.threshold.OOB"] else
prob_threshold_def <- NULL
for (obsSet in c("trn", "new")) {
obs_df <- switch(obsSet, all = glb_allobs_df, trn = glb_trnobs_df, new = glb_newobs_df)
obs_df <- glb_get_predictions(obs_df, mdl_id = glb_fin_mdl_id,
rsp_var_out = glb_rsp_var_out, prob_threshold_def = prob_threshold_def)
if (obsSet == "all") glb_allobs_df <- obs_df else
if (obsSet == "trn") glb_trnobs_df <- obs_df else
if (obsSet == "new") glb_newobs_df <- obs_df
}## Warning in glb_get_predictions(obs_df, mdl_id = glb_fin_mdl_id, rsp_var_out
## = glb_rsp_var_out, : Using default probability threshold: 0.3## Warning in glb_get_predictions(obs_df, mdl_id = glb_fin_mdl_id, rsp_var_out
## = glb_rsp_var_out, : Using default probability threshold: 0.3rm(obs_df)
glb_allobs_df <- orderBy(reformulate(glb_id_var), myrbind_df(glb_trnobs_df, glb_newobs_df))
glb_analytics_diag_plots(obs_df = glb_newobs_df, mdl_id = glb_fin_mdl_id,
prob_threshold = prob_threshold_def)## Warning in glb_analytics_diag_plots(obs_df = glb_newobs_df, mdl_id =
## glb_fin_mdl_id, : Limiting important feature scatter plots to 5 out of 32## Warning: Removed 798 rows containing missing values (geom_point).## Warning: Removed 798 rows containing missing values (geom_point).
## Warning: Removed 798 rows containing missing values (geom_point).## Warning: Removed 798 rows containing missing values (geom_point).
## Warning: Removed 798 rows containing missing values (geom_point).## Warning: Removed 798 rows containing missing values (geom_point).
## Warning: Removed 798 rows containing missing values (geom_point).## Warning: Removed 798 rows containing missing values (geom_point).
## Warning: Removed 798 rows containing missing values (geom_point).## Warning: Removed 798 rows containing missing values (geom_point).
## [1] "Min/Max Boundaries: "
## UniqueID sold.fctr sold.fctr.predict.Final.glmnet.prob
## 1851 11862 <NA> 0.4737074
## 1854 11865 <NA> 0.8390796
## sold.fctr.predict.Final.glmnet sold.fctr.predict.Final.glmnet.err
## 1851 Y NA
## 1854 Y NA
## sold.fctr.predict.Final.glmnet.err.abs
## 1851 NA
## 1854 NA
## sold.fctr.predict.Final.glmnet.accurate
## 1851 NA
## 1854 NA
## sold.fctr.predict.Final.glmnet.error .label
## 1851 0 11862
## 1854 0 11865
## [1] "Inaccurate: "
## UniqueID sold.fctr sold.fctr.predict.Final.glmnet.prob
## NA NA <NA> NA
## NA.1 NA <NA> NA
## NA.2 NA <NA> NA
## NA.3 NA <NA> NA
## NA.4 NA <NA> NA
## NA.5 NA <NA> NA
## sold.fctr.predict.Final.glmnet sold.fctr.predict.Final.glmnet.err
## NA <NA> NA
## NA.1 <NA> NA
## NA.2 <NA> NA
## NA.3 <NA> NA
## NA.4 <NA> NA
## NA.5 <NA> NA
## sold.fctr.predict.Final.glmnet.err.abs
## NA NA
## NA.1 NA
## NA.2 NA
## NA.3 NA
## NA.4 NA
## NA.5 NA
## sold.fctr.predict.Final.glmnet.accurate
## NA NA
## NA.1 NA
## NA.2 NA
## NA.3 NA
## NA.4 NA
## NA.5 NA
## sold.fctr.predict.Final.glmnet.error
## NA NA
## NA.1 NA
## NA.2 NA
## NA.3 NA
## NA.4 NA
## NA.5 NA
## UniqueID sold.fctr sold.fctr.predict.Final.glmnet.prob
## NA.6 NA <NA> NA
## NA.227 NA <NA> NA
## NA.241 NA <NA> NA
## NA.380 NA <NA> NA
## NA.446 NA <NA> NA
## NA.507 NA <NA> NA
## sold.fctr.predict.Final.glmnet sold.fctr.predict.Final.glmnet.err
## NA.6 <NA> NA
## NA.227 <NA> NA
## NA.241 <NA> NA
## NA.380 <NA> NA
## NA.446 <NA> NA
## NA.507 <NA> NA
## sold.fctr.predict.Final.glmnet.err.abs
## NA.6 NA
## NA.227 NA
## NA.241 NA
## NA.380 NA
## NA.446 NA
## NA.507 NA
## sold.fctr.predict.Final.glmnet.accurate
## NA.6 NA
## NA.227 NA
## NA.241 NA
## NA.380 NA
## NA.446 NA
## NA.507 NA
## sold.fctr.predict.Final.glmnet.error
## NA.6 NA
## NA.227 NA
## NA.241 NA
## NA.380 NA
## NA.446 NA
## NA.507 NA
## UniqueID sold.fctr sold.fctr.predict.Final.glmnet.prob
## NA.792 NA <NA> NA
## NA.793 NA <NA> NA
## NA.794 NA <NA> NA
## NA.795 NA <NA> NA
## NA.796 NA <NA> NA
## NA.797 NA <NA> NA
## sold.fctr.predict.Final.glmnet sold.fctr.predict.Final.glmnet.err
## NA.792 <NA> NA
## NA.793 <NA> NA
## NA.794 <NA> NA
## NA.795 <NA> NA
## NA.796 <NA> NA
## NA.797 <NA> NA
## sold.fctr.predict.Final.glmnet.err.abs
## NA.792 NA
## NA.793 NA
## NA.794 NA
## NA.795 NA
## NA.796 NA
## NA.797 NA
## sold.fctr.predict.Final.glmnet.accurate
## NA.792 NA
## NA.793 NA
## NA.794 NA
## NA.795 NA
## NA.796 NA
## NA.797 NA
## sold.fctr.predict.Final.glmnet.error
## NA.792 NA
## NA.793 NA
## NA.794 NA
## NA.795 NA
## NA.796 NA
## NA.797 NA## Warning: Removed 798 rows containing missing values (geom_point).
if (is.null(glb_out_obs)) obs_df <- glb_newobs_df else
obs_df <- switch(glb_out_obs,
all = glb_allobs_df, trn = glb_trnobs_df, new = glb_newobs_df)
require(stringr)
if (glb_is_classification && glb_is_binomial) {
# submit_df <- glb_newobs_df[, c(glb_id_var,
# paste0(glb_rsp_var_out, glb_fin_mdl_id, ".prob"))]
# names(submit_df)[2] <- "Probability1"
obsout_df <- obs_df[, glb_id_var, FALSE]
for (clmn in names(glb_out_vars_lst))
if (!grepl("^%<d-%", glb_out_vars_lst[[clmn]]))
obsout_df[, clmn] <- obs_df[, glb_out_vars_lst[[clmn]]] else {
feat <- str_trim(unlist(strsplit(glb_out_vars_lst[[clmn]], "%<d-%"))[2])
obsout_df[, clmn] <- obs_df[, eval(parse(text = feat))]
}
glb_force_prediction_lst <- list()
glb_force_prediction_lst[["0"]] <- c(11885, 11907, 11932, 11943,
12050, 12115, 12171,
12253, 12285, 12367, 12388, 12399,
12585)
for (obs_id in glb_force_prediction_lst[["0"]]) {
if (sum(glb_allobs_df[, glb_id_var] == obs_id) == 0)
next
if (is.na(glb_allobs_df[glb_allobs_df[, glb_id_var] == obs_id, ".grpid"]))
stop(".grpid is NA")
# submit_df[submit_df[, glb_id_var] == obs_id, "Probability1"] <-
# max(0, submit_df[submit_df[, glb_id_var] == obs_id, "Probability1"] - 0.5)
}
glb_force_prediction_lst[["1"]] <- c(11871, 11875, 11886,
11913, 11931, 11937, 11967, 11982, 11990, 11991, 11994, 11999,
12000, 12002, 12004, 12018, 12021, 12065, 12072,
12111, 12114, 12126, 12134, 12152, 12172,
12213, 12214, 12233, 12265, 12278, 12299,
12446, 12491,
12505, 12576, 12608, 12630)
for (obs_id in glb_force_prediction_lst[["1"]]) {
if (sum(glb_allobs_df[, glb_id_var] == obs_id) == 0)
next
if (is.na(glb_allobs_df[glb_allobs_df[, glb_id_var] == obs_id, ".grpid"]))
stop(".grpid is NA")
# submit_df[submit_df[, glb_id_var] == obs_id, "Probability1"] <-
# min(0.9999, submit_df[submit_df[, glb_id_var] == obs_id, "Probability1"] + 0.5)
}
rsp_var_out <- paste0(glb_rsp_var_out, glb_fin_mdl_id)
for (obs_id in glb_newobs_df[!is.na(glb_newobs_df[, rsp_var_out]) &
(glb_newobs_df[, rsp_var_out] == "Y") &
(glb_newobs_df[ , "startprice"] > 675), "UniqueID"]) {
# submit_df[submit_df[, glb_id_var] == obs_id, "Probability1"] <-
# max(0, submit_df[submit_df[, glb_id_var] == obs_id, "Probability1"] - 0.5)
}
} else {
# submit_df <- glb_newobs_df[, c(glb_id_var,
# paste0(glb_rsp_var_out, glb_fin_mdl_id))]
obsout_df <- obs_df[, glb_id_var, FALSE]
for (clmn in names(glb_out_vars_lst))
if (!grepl("^%<d-%", glb_out_vars_lst[[clmn]]))
obsout_df[, clmn] <- obs_df[, glb_out_vars_lst[[clmn]]] else {
feat <- str_trim(unlist(strsplit(glb_out_vars_lst[[clmn]], "%<d-%"))[2])
obsout_df[, clmn] <- obs_df[, eval(parse(text=feat))]
}
}
if (glb_is_classification) {
rsp_var_out <- paste0(glb_rsp_var_out, glb_fin_mdl_id)
tmp_newobs_df <- subset(glb_newobs_df[, c(glb_id_var, ".grpid", rsp_var_out)],
!is.na(.grpid))
tmp_newobs_df <- merge(tmp_newobs_df, dupgrps_df, by = ".grpid", all.x = TRUE)
tmp_newobs_df <- merge(tmp_newobs_df, obsout_df, by = glb_id_var, all.x = TRUE)
tmp_newobs_df$.err <-
((tmp_newobs_df$Probability1 > 0.5) & (tmp_newobs_df$sold.0 > 0) |
(tmp_newobs_df$Probability1 < 0.5) & (tmp_newobs_df$sold.1 > 0))
tmp_newobs_df <- orderBy(~UniqueID, subset(tmp_newobs_df, .err == TRUE))
print(sprintf("Prediction errors in duplicates: %d", nrow(tmp_newobs_df)))
print(tmp_newobs_df)
# if (nrow(tmp_newobs_df) > 0)
# stop("check Prediction errors in duplicates")
#print(dupobs_df[dupobs_df$.grpid == 26, ])
tmp_newobs_df <- cbind(glb_newobs_df, obsout_df[, "Probability1", FALSE])
# if (max(glb_newobs_df[!is.na(glb_newobs_df[, rsp_var_out]) &
# (tmp_newobs_df[, "Probability1"] >= 0.5), "startprice"]) >
# max(glb_allobs_df[!is.na(glb_allobs_df[, glb_rsp_var]) &
# (glb_allobs_df[, glb_rsp_var] == "Y"), "startprice"]))
# stop("startprice for some +ve predictions > 675")
# Check predictions that are outside of data ranges
#stop(here")
require(stringr)
tmp_feats_df <- subset(glb_feats_df,
!nzv &
(exclude.as.feat != 1) &
!grepl(".fctr", id, fixed=TRUE))[, "id", FALSE]
ranges_all_df <- glb_allobs_df[, tmp_feats_df$id] %>%
dplyr::summarise_each(funs(min(., na.rm=TRUE),
max(., na.rm=TRUE))) %>%
tidyr::gather() %>%
dplyr::mutate(id=str_sub(key, 1, -5),
stat=str_sub(key, -3)) %>%
dplyr::select(-key) %>%
tidyr::spread(stat, value)
# sav_ranges_trn_df <- ranges_trn_df; all.equal(sav_ranges_trn_df, ranges_trn_df)
# sav_ranges_new_df <- ranges_new_df; all.equal(sav_ranges_new_df, ranges_new_df)
get_ranges_df <- function(obs_df, feats, class_var) {
require(tidyr)
ranges_df <- obs_df[, c(class_var, feats)] %>%
dplyr::group_by_(class_var) %>%
dplyr::summarise_each(funs(min(., na.rm=TRUE),
max(., na.rm=TRUE))) %>%
tidyr::gather(key, value, -1) %>%
mutate(id=str_sub(key, 1, -5),
stat.vname=paste0(str_sub(key, -3), ".", class_var)) %>%
unite_("stat.class", c("stat.vname", class_var), sep=".") %>%
dplyr::select(-key) %>%
spread(stat.class, value)
return(ranges_df)
}
rsp_var_out_OOB <- paste0(glb_rsp_var_out, glb_sel_mdl_id)
rsp_var_out_new <- paste0(glb_rsp_var_out, glb_fin_mdl_id)
ranges_trn_df <- get_ranges_df(obs_df=glb_trnobs_df, feats=tmp_feats_df$id,
class_var=glb_rsp_var)
ranges_fit_df <- get_ranges_df(obs_df=glb_fitobs_df, feats=tmp_feats_df$id,
class_var=glb_rsp_var)
ranges_OOB_df <- get_ranges_df(obs_df=glb_OOBobs_df, feats=tmp_feats_df$id,
class_var=rsp_var_out_OOB)
ranges_new_df <- get_ranges_df(obs_df=glb_newobs_df, feats=tmp_feats_df$id,
class_var=rsp_var_out_new)
for (obsset in c("OOB", "new")) {
if (obsset == "OOB") {
ranges_ref_df <- ranges_fit_df; obs_df <- glb_OOBobs_df;
rsp_var_out_obs <- rsp_var_out_OOB; sprintf_pfx <- "OOBobs";
} else {
ranges_ref_df <- ranges_trn_df; obs_df <- glb_newobs_df;
rsp_var_out_obs <- rsp_var_out_new; sprintf_pfx <- "newobs";
}
plt_feats_df <- glb_feats_df %>%
merge(ranges_all_df, all=TRUE) %>%
merge(ranges_ref_df, all=TRUE) %>%
merge(ranges_OOB_df, all=TRUE) %>%
merge(ranges_new_df, all=TRUE) %>%
subset(!is.na(min) & (id != ".rnorm"))
row.names(plt_feats_df) <- plt_feats_df$id
range_outlier_ids <- c(NULL)
for (clss in unique(obs_df[, rsp_var_out_obs])) {
for (stat in c("min", "max")) {
if (stat == "min") {
dsp_feats <- plt_feats_df[
which(plt_feats_df[, paste("min", rsp_var_out_obs, clss, sep=".")] <
plt_feats_df[, paste("min", glb_rsp_var, clss, sep=".")]), "id"]
} else {
dsp_feats <- plt_feats_df[
which(plt_feats_df[, paste("max", rsp_var_out_obs, clss, sep=".")] >
plt_feats_df[, paste("max", glb_rsp_var, clss, sep=".")]), "id"]
}
if (length(dsp_feats) > 0) {
ths_ids <- c(NULL)
for (feat in dsp_feats) {
if (stat == "min") {
ths_ids <- union(ths_ids,
obs_df[(obs_df[, rsp_var_out_obs] == clss) &
(obs_df[, feat] <
plt_feats_df[plt_feats_df$id == feat, paste("min", glb_rsp_var, clss, sep=".")]),
glb_id_var])
} else {
ths_ids <- union(ths_ids,
obs_df[(obs_df[, rsp_var_out_obs] == clss) &
(obs_df[, feat] >
plt_feats_df[plt_feats_df$id == feat, paste("max", glb_rsp_var, clss, sep=".")]),
glb_id_var])
}
}
tmp_obs_df <- obs_df[obs_df[, glb_id_var] %in% ths_ids,
c(glb_id_var, rsp_var_out_obs, dsp_feats)]
if (stat == "min") {
print(sprintf("%s %s %s: min < min of Train range: %d",
sprintf_pfx, rsp_var_out_obs, clss, nrow(tmp_obs_df)))
} else {
print(sprintf("%s %s %s: max > max of Train range: %d",
sprintf_pfx, rsp_var_out_obs, clss, nrow(tmp_obs_df)))
}
myprint_df(tmp_obs_df)
print(subset(plt_feats_df, id %in% dsp_feats))
range_outlier_ids <- union(range_outlier_ids, ths_ids)
}
}
}
print(sprintf("%s total range outliers: %d", sprintf_pfx, length(range_outlier_ids)))
}
} ## [1] "Prediction errors in duplicates: 15"
## UniqueID .grpid sold.fctr.predict.Final.glmnet sold.0 sold.1 sold.NA
## 4 11886 134 Y 0 1 1
## 9 11931 53 Y 0 2 2
## 23 11994 101 Y 0 1 1
## 24 11999 115 Y 0 1 1
## 42 12111 128 Y 0 1 1
## 44 12115 56 Y 1 0 1
## 59 12233 30 Y 0 1 1
## 68 12285 55 Y 1 0 2
## 69 12299 53 Y 0 2 2
## 72 12367 55 Y 1 0 2
## 76 12446 103 N 0 1 1
## 80 12505 54 Y 0 1 1
## 81 12576 93 N 0 1 1
## 82 12585 112 Y 1 0 1
## 87 12630 76 Y 0 2 1
## .freq Probability1 .err
## 4 2 0.3527705 TRUE
## 9 4 0.4356090 TRUE
## 23 2 0.4130907 TRUE
## 24 2 0.4077973 TRUE
## 42 2 0.4360890 TRUE
## 44 2 0.5274436 TRUE
## 59 2 0.3347090 TRUE
## 68 3 0.5633672 TRUE
## 69 4 0.4356090 TRUE
## 72 3 0.5633672 TRUE
## 76 2 0.2123043 TRUE
## 80 2 0.4246938 TRUE
## 81 2 0.2095837 TRUE
## 82 2 0.5577166 TRUE
## 87 3 0.3397270 TRUE
## [1] "OOBobs sold.fctr.predict.All.X.glmnet N: max > max of Train range: 6"
## UniqueID sold.fctr.predict.All.X.glmnet D.chrs.uppr.n.log
## 1294 11301 N 4.465908
## 1117 11123 N 4.343805
## 1633 11642 N 2.708050
## 1806 11816 N 4.382027
## 1767 11776 N 3.850148
## 1810 11820 N 4.382027
## D.terms.post.stem.n.log D.weight.post.stem.sum D.weight.post.stop.sum
## 1294 2.564949 4.522716 5.268255
## 1117 2.639057 6.824967 7.020585
## 1633 1.098612 6.917242 6.917242
## 1806 2.995732 4.193223 4.241549
## 1767 2.302585 7.257702 7.278648
## 1810 2.995732 4.193223 4.241549
## D.weight.sum D.wrds.unq.n.log
## 1294 4.522716 2.564949
## 1117 6.824967 2.639057
## 1633 6.917242 1.098612
## 1806 4.193223 2.995732
## 1767 7.257702 2.302585
## 1810 4.193223 2.995732
## id cor.y
## D.chrs.uppr.n.log D.chrs.uppr.n.log -0.05449161
## D.terms.post.stem.n.log D.terms.post.stem.n.log -0.06222063
## D.weight.post.stem.sum D.weight.post.stem.sum -0.04710544
## D.weight.post.stop.sum D.weight.post.stop.sum -0.04500921
## D.weight.sum D.weight.sum -0.04710544
## D.wrds.unq.n.log D.wrds.unq.n.log -0.06222063
## exclude.as.feat cor.y.abs cor.high.X
## D.chrs.uppr.n.log FALSE 0.05449161 D.chrs.n.log
## D.terms.post.stem.n.log FALSE 0.06222063 D.terms.post.stop.n.log
## D.weight.post.stem.sum FALSE 0.04710544 D.terms.post.stop.n.log
## D.weight.post.stop.sum FALSE 0.04500921 D.weight.post.stem.sum
## D.weight.sum FALSE 0.04710544 D.weight.post.stem.sum
## D.wrds.unq.n.log FALSE 0.06222063 D.terms.post.stem.n.log
## freqRatio percentUnique zeroVar nzv
## D.chrs.uppr.n.log 15.66176 4.486486 FALSE FALSE
## D.terms.post.stem.n.log 11.71429 1.135135 FALSE FALSE
## D.weight.post.stem.sum 62.70588 34.648649 FALSE FALSE
## D.weight.post.stop.sum 62.70588 34.756757 FALSE FALSE
## D.weight.sum 62.70588 34.648649 FALSE FALSE
## D.wrds.unq.n.log 11.71429 1.135135 FALSE FALSE
## is.cor.y.abs.low interaction.feat
## D.chrs.uppr.n.log FALSE <NA>
## D.terms.post.stem.n.log FALSE <NA>
## D.weight.post.stem.sum FALSE <NA>
## D.weight.post.stop.sum FALSE <NA>
## D.weight.sum FALSE <NA>
## D.wrds.unq.n.log FALSE <NA>
## shapiro.test.p.value rsp_var_raw id_var rsp_var
## D.chrs.uppr.n.log 4.550442e-50 FALSE NA NA
## D.terms.post.stem.n.log 1.386439e-48 FALSE NA NA
## D.weight.post.stem.sum 2.197748e-48 FALSE NA NA
## D.weight.post.stop.sum 2.648991e-48 FALSE NA NA
## D.weight.sum 2.197748e-48 FALSE NA NA
## D.wrds.unq.n.log 1.386439e-48 FALSE NA NA
## max min max.sold.fctr.N max.sold.fctr.Y
## D.chrs.uppr.n.log 4.465908 0 4.454347 4.454347
## D.terms.post.stem.n.log 3.044522 0 2.944439 2.995732
## D.weight.post.stem.sum 8.048759 0 6.702992 7.048759
## D.weight.post.stop.sum 8.563332 0 7.202321 7.048759
## D.weight.sum 8.048759 0 6.702992 7.048759
## D.wrds.unq.n.log 3.044522 0 2.944439 2.995732
## min.sold.fctr.N min.sold.fctr.Y
## D.chrs.uppr.n.log 0 0
## D.terms.post.stem.n.log 0 0
## D.weight.post.stem.sum 0 0
## D.weight.post.stop.sum 0 0
## D.weight.sum 0 0
## D.wrds.unq.n.log 0 0
## max.sold.fctr.predict.All.X.glmnet.N
## D.chrs.uppr.n.log 4.465908
## D.terms.post.stem.n.log 2.995732
## D.weight.post.stem.sum 7.257702
## D.weight.post.stop.sum 7.278648
## D.weight.sum 7.257702
## D.wrds.unq.n.log 2.995732
## max.sold.fctr.predict.All.X.glmnet.Y
## D.chrs.uppr.n.log 4.442651
## D.terms.post.stem.n.log 3.044522
## D.weight.post.stem.sum 8.048759
## D.weight.post.stop.sum 8.563332
## D.weight.sum 8.048759
## D.wrds.unq.n.log 3.044522
## min.sold.fctr.predict.All.X.glmnet.N
## D.chrs.uppr.n.log 0
## D.terms.post.stem.n.log 0
## D.weight.post.stem.sum 0
## D.weight.post.stop.sum 0
## D.weight.sum 0
## D.wrds.unq.n.log 0
## min.sold.fctr.predict.All.X.glmnet.Y
## D.chrs.uppr.n.log 0
## D.terms.post.stem.n.log 0
## D.weight.post.stem.sum 0
## D.weight.post.stop.sum 0
## D.weight.sum 0
## D.wrds.unq.n.log 0
## max.sold.fctr.predict.Final.glmnet.N
## D.chrs.uppr.n.log 4.465908
## D.terms.post.stem.n.log 2.944439
## D.weight.post.stem.sum 6.605876
## D.weight.post.stop.sum 6.682933
## D.weight.sum 6.605876
## D.wrds.unq.n.log 2.944439
## max.sold.fctr.predict.Final.glmnet.Y
## D.chrs.uppr.n.log 4.442651
## D.terms.post.stem.n.log 2.944439
## D.weight.post.stem.sum 6.253320
## D.weight.post.stop.sum 6.851448
## D.weight.sum 6.253320
## D.wrds.unq.n.log 2.944439
## min.sold.fctr.predict.Final.glmnet.N
## D.chrs.uppr.n.log 0
## D.terms.post.stem.n.log 0
## D.weight.post.stem.sum 0
## D.weight.post.stop.sum 0
## D.weight.sum 0
## D.wrds.unq.n.log 0
## min.sold.fctr.predict.Final.glmnet.Y
## D.chrs.uppr.n.log 0
## D.terms.post.stem.n.log 0
## D.weight.post.stem.sum 0
## D.weight.post.stop.sum 0
## D.weight.sum 0
## D.wrds.unq.n.log 0
## [1] "OOBobs sold.fctr.predict.All.X.glmnet Y: min < min of Train range: 3"
## UniqueID sold.fctr.predict.All.X.glmnet D.ratio.wrds.stop.n.wrds.n
## 915 10918 Y 0.04347826
## 1594 11603 Y 0.40000000
## 227 10227 Y 0.30000000
## D.weight.sum.stem.stop.Ratio
## 915 0.9756676
## 1594 0.7143207
## 227 0.7741412
## id cor.y
## D.ratio.wrds.stop.n.wrds.n D.ratio.wrds.stop.n.wrds.n 0.059547328
## D.weight.sum.stem.stop.Ratio D.weight.sum.stem.stop.Ratio 0.002381527
## exclude.as.feat cor.y.abs
## D.ratio.wrds.stop.n.wrds.n FALSE 0.059547328
## D.weight.sum.stem.stop.Ratio FALSE 0.002381527
## cor.high.X freqRatio
## D.ratio.wrds.stop.n.wrds.n D.terms.post.stop.n.log 18.06780
## D.weight.sum.stem.stop.Ratio <NA> 64.47059
## percentUnique zeroVar nzv is.cor.y.abs.low
## D.ratio.wrds.stop.n.wrds.n 4.27027 FALSE FALSE FALSE
## D.weight.sum.stem.stop.Ratio 33.56757 FALSE FALSE FALSE
## interaction.feat shapiro.test.p.value
## D.ratio.wrds.stop.n.wrds.n <NA> 5.138767e-49
## D.weight.sum.stem.stop.Ratio <NA> 7.686585e-54
## rsp_var_raw id_var rsp_var max min
## D.ratio.wrds.stop.n.wrds.n FALSE NA NA 1 0.04347826
## D.weight.sum.stem.stop.Ratio FALSE NA NA 1 0.71432071
## max.sold.fctr.N max.sold.fctr.Y
## D.ratio.wrds.stop.n.wrds.n 1 1
## D.weight.sum.stem.stop.Ratio 1 1
## min.sold.fctr.N min.sold.fctr.Y
## D.ratio.wrds.stop.n.wrds.n 0.04347826 0.04545455
## D.weight.sum.stem.stop.Ratio 0.76580062 0.81266745
## max.sold.fctr.predict.All.X.glmnet.N
## D.ratio.wrds.stop.n.wrds.n 1
## D.weight.sum.stem.stop.Ratio 1
## max.sold.fctr.predict.All.X.glmnet.Y
## D.ratio.wrds.stop.n.wrds.n 1
## D.weight.sum.stem.stop.Ratio 1
## min.sold.fctr.predict.All.X.glmnet.N
## D.ratio.wrds.stop.n.wrds.n 0.04761905
## D.weight.sum.stem.stop.Ratio 0.80128574
## min.sold.fctr.predict.All.X.glmnet.Y
## D.ratio.wrds.stop.n.wrds.n 0.04347826
## D.weight.sum.stem.stop.Ratio 0.71432071
## max.sold.fctr.predict.Final.glmnet.N
## D.ratio.wrds.stop.n.wrds.n 1
## D.weight.sum.stem.stop.Ratio 1
## max.sold.fctr.predict.Final.glmnet.Y
## D.ratio.wrds.stop.n.wrds.n 1
## D.weight.sum.stem.stop.Ratio 1
## min.sold.fctr.predict.Final.glmnet.N
## D.ratio.wrds.stop.n.wrds.n 0.0500000
## D.weight.sum.stem.stop.Ratio 0.8220163
## min.sold.fctr.predict.Final.glmnet.Y
## D.ratio.wrds.stop.n.wrds.n 0.0500000
## D.weight.sum.stem.stop.Ratio 0.7284632
## [1] "OOBobs sold.fctr.predict.All.X.glmnet Y: max > max of Train range: 6"
## UniqueID sold.fctr.predict.All.X.glmnet D.terms.post.stem.n.log
## 873 10876 Y 1.3862944
## 1612 11621 Y 2.3025851
## 915 10918 Y 3.0445224
## 1594 11603 Y 1.3862944
## 119 10119 Y 1.6094379
## 812 10814 Y 0.6931472
## D.terms.post.stop.n.log D.weight.post.stem.sum D.weight.post.stop.sum
## 873 1.3862944 7.258891 7.484915
## 1612 2.3025851 4.333960 4.498634
## 915 3.0445224 4.623058 4.738354
## 1594 1.3862944 5.066153 7.092267
## 119 1.6094379 7.332280 7.332280
## 812 0.6931472 8.048759 8.563332
## D.weight.sum D.wrds.stop.n.log D.wrds.unq.n.log
## 873 7.258891 0.6931472 1.3862944
## 1612 4.333960 2.3025851 2.3025851
## 915 4.623058 0.0000000 3.0445224
## 1594 5.066153 0.6931472 1.3862944
## 119 7.332280 0.6931472 1.6094379
## 812 8.048759 1.0986123 0.6931472
## id cor.y
## D.terms.post.stem.n.log D.terms.post.stem.n.log -0.062220631
## D.terms.post.stop.n.log D.terms.post.stop.n.log -0.062526435
## D.weight.post.stem.sum D.weight.post.stem.sum -0.047105442
## D.weight.post.stop.sum D.weight.post.stop.sum -0.045009208
## D.weight.sum D.weight.sum -0.047105442
## D.wrds.stop.n.log D.wrds.stop.n.log 0.003851183
## D.wrds.unq.n.log D.wrds.unq.n.log -0.062220631
## exclude.as.feat cor.y.abs
## D.terms.post.stem.n.log FALSE 0.062220631
## D.terms.post.stop.n.log FALSE 0.062526435
## D.weight.post.stem.sum FALSE 0.047105442
## D.weight.post.stop.sum FALSE 0.045009208
## D.weight.sum FALSE 0.047105442
## D.wrds.stop.n.log FALSE 0.003851183
## D.wrds.unq.n.log FALSE 0.062220631
## cor.high.X freqRatio percentUnique
## D.terms.post.stem.n.log D.terms.post.stop.n.log 11.714286 1.1351351
## D.terms.post.stop.n.log <NA> 13.325000 1.1351351
## D.weight.post.stem.sum D.terms.post.stop.n.log 62.705882 34.6486486
## D.weight.post.stop.sum D.weight.post.stem.sum 62.705882 34.7567568
## D.weight.sum D.weight.post.stem.sum 62.705882 34.6486486
## D.wrds.stop.n.log <NA> 8.732484 0.5405405
## D.wrds.unq.n.log D.terms.post.stem.n.log 11.714286 1.1351351
## zeroVar nzv is.cor.y.abs.low interaction.feat
## D.terms.post.stem.n.log FALSE FALSE FALSE <NA>
## D.terms.post.stop.n.log FALSE FALSE FALSE <NA>
## D.weight.post.stem.sum FALSE FALSE FALSE <NA>
## D.weight.post.stop.sum FALSE FALSE FALSE <NA>
## D.weight.sum FALSE FALSE FALSE <NA>
## D.wrds.stop.n.log FALSE FALSE FALSE <NA>
## D.wrds.unq.n.log FALSE FALSE FALSE <NA>
## shapiro.test.p.value rsp_var_raw id_var rsp_var
## D.terms.post.stem.n.log 1.386439e-48 FALSE NA NA
## D.terms.post.stop.n.log 1.391188e-48 FALSE NA NA
## D.weight.post.stem.sum 2.197748e-48 FALSE NA NA
## D.weight.post.stop.sum 2.648991e-48 FALSE NA NA
## D.weight.sum 2.197748e-48 FALSE NA NA
## D.wrds.stop.n.log 4.444633e-54 FALSE NA NA
## D.wrds.unq.n.log 1.386439e-48 FALSE NA NA
## max min max.sold.fctr.N max.sold.fctr.Y
## D.terms.post.stem.n.log 3.044522 0 2.944439 2.995732
## D.terms.post.stop.n.log 3.044522 0 2.995732 2.995732
## D.weight.post.stem.sum 8.048759 0 6.702992 7.048759
## D.weight.post.stop.sum 8.563332 0 7.202321 7.048759
## D.weight.sum 8.048759 0 6.702992 7.048759
## D.wrds.stop.n.log 2.564949 0 2.302585 2.197225
## D.wrds.unq.n.log 3.044522 0 2.944439 2.995732
## min.sold.fctr.N min.sold.fctr.Y
## D.terms.post.stem.n.log 0 0
## D.terms.post.stop.n.log 0 0
## D.weight.post.stem.sum 0 0
## D.weight.post.stop.sum 0 0
## D.weight.sum 0 0
## D.wrds.stop.n.log 0 0
## D.wrds.unq.n.log 0 0
## max.sold.fctr.predict.All.X.glmnet.N
## D.terms.post.stem.n.log 2.995732
## D.terms.post.stop.n.log 2.995732
## D.weight.post.stem.sum 7.257702
## D.weight.post.stop.sum 7.278648
## D.weight.sum 7.257702
## D.wrds.stop.n.log 2.302585
## D.wrds.unq.n.log 2.995732
## max.sold.fctr.predict.All.X.glmnet.Y
## D.terms.post.stem.n.log 3.044522
## D.terms.post.stop.n.log 3.044522
## D.weight.post.stem.sum 8.048759
## D.weight.post.stop.sum 8.563332
## D.weight.sum 8.048759
## D.wrds.stop.n.log 2.302585
## D.wrds.unq.n.log 3.044522
## min.sold.fctr.predict.All.X.glmnet.N
## D.terms.post.stem.n.log 0
## D.terms.post.stop.n.log 0
## D.weight.post.stem.sum 0
## D.weight.post.stop.sum 0
## D.weight.sum 0
## D.wrds.stop.n.log 0
## D.wrds.unq.n.log 0
## min.sold.fctr.predict.All.X.glmnet.Y
## D.terms.post.stem.n.log 0
## D.terms.post.stop.n.log 0
## D.weight.post.stem.sum 0
## D.weight.post.stop.sum 0
## D.weight.sum 0
## D.wrds.stop.n.log 0
## D.wrds.unq.n.log 0
## max.sold.fctr.predict.Final.glmnet.N
## D.terms.post.stem.n.log 2.944439
## D.terms.post.stop.n.log 2.944439
## D.weight.post.stem.sum 6.605876
## D.weight.post.stop.sum 6.682933
## D.weight.sum 6.605876
## D.wrds.stop.n.log 2.397895
## D.wrds.unq.n.log 2.944439
## max.sold.fctr.predict.Final.glmnet.Y
## D.terms.post.stem.n.log 2.944439
## D.terms.post.stop.n.log 2.944439
## D.weight.post.stem.sum 6.253320
## D.weight.post.stop.sum 6.851448
## D.weight.sum 6.253320
## D.wrds.stop.n.log 2.564949
## D.wrds.unq.n.log 2.944439
## min.sold.fctr.predict.Final.glmnet.N
## D.terms.post.stem.n.log 0
## D.terms.post.stop.n.log 0
## D.weight.post.stem.sum 0
## D.weight.post.stop.sum 0
## D.weight.sum 0
## D.wrds.stop.n.log 0
## D.wrds.unq.n.log 0
## min.sold.fctr.predict.Final.glmnet.Y
## D.terms.post.stem.n.log 0
## D.terms.post.stop.n.log 0
## D.weight.post.stem.sum 0
## D.weight.post.stop.sum 0
## D.weight.sum 0
## D.wrds.stop.n.log 0
## D.wrds.unq.n.log 0
## [1] "OOBobs total range outliers: 13"
## [1] "newobs sold.fctr.predict.Final.glmnet Y: max > max of Train range: 4"
## UniqueID sold.fctr.predict.Final.glmnet D.chrs.n.log
## 2020 12031 Y 4.663439
## 2048 12059 Y 3.737670
## 2076 12087 Y 4.663439
## 2329 12340 Y 4.624973
## D.chrs.pnct13.n.log D.wrds.stop.n.log
## 2020 0.6931472 1.386294
## 2048 2.7080502 0.000000
## 2076 0.0000000 1.098612
## 2329 0.0000000 2.564949
## id cor.y exclude.as.feat
## D.chrs.n.log D.chrs.n.log -0.055576665 FALSE
## D.chrs.pnct13.n.log D.chrs.pnct13.n.log -0.032095777 FALSE
## D.wrds.stop.n.log D.wrds.stop.n.log 0.003851183 FALSE
## cor.y.abs cor.high.X freqRatio
## D.chrs.n.log 0.055576665 D.terms.post.stop.n.log 13.455696
## D.chrs.pnct13.n.log 0.032095777 D.terms.post.stop.n.log 5.272374
## D.wrds.stop.n.log 0.003851183 <NA> 8.732484
## percentUnique zeroVar nzv is.cor.y.abs.low
## D.chrs.n.log 5.3513514 FALSE FALSE FALSE
## D.chrs.pnct13.n.log 0.4864865 FALSE FALSE FALSE
## D.wrds.stop.n.log 0.5405405 FALSE FALSE FALSE
## interaction.feat shapiro.test.p.value rsp_var_raw
## D.chrs.n.log <NA> 4.162678e-50 FALSE
## D.chrs.pnct13.n.log <NA> 1.157759e-53 FALSE
## D.wrds.stop.n.log <NA> 4.444633e-54 FALSE
## id_var rsp_var max min max.sold.fctr.N
## D.chrs.n.log NA NA 4.682131 0 4.682131
## D.chrs.pnct13.n.log NA NA 2.708050 0 1.945910
## D.wrds.stop.n.log NA NA 2.564949 0 2.302585
## max.sold.fctr.Y min.sold.fctr.N min.sold.fctr.Y
## D.chrs.n.log 4.644391 0 0
## D.chrs.pnct13.n.log 2.197225 0 0
## D.wrds.stop.n.log 2.302585 0 0
## max.sold.fctr.predict.All.X.glmnet.N
## D.chrs.n.log 4.644391
## D.chrs.pnct13.n.log 1.609438
## D.wrds.stop.n.log 2.302585
## max.sold.fctr.predict.All.X.glmnet.Y
## D.chrs.n.log 4.624973
## D.chrs.pnct13.n.log 1.791759
## D.wrds.stop.n.log 2.302585
## min.sold.fctr.predict.All.X.glmnet.N
## D.chrs.n.log 0
## D.chrs.pnct13.n.log 0
## D.wrds.stop.n.log 0
## min.sold.fctr.predict.All.X.glmnet.Y
## D.chrs.n.log 0
## D.chrs.pnct13.n.log 0
## D.wrds.stop.n.log 0
## max.sold.fctr.predict.Final.glmnet.N
## D.chrs.n.log 4.663439
## D.chrs.pnct13.n.log 1.791759
## D.wrds.stop.n.log 2.397895
## max.sold.fctr.predict.Final.glmnet.Y
## D.chrs.n.log 4.663439
## D.chrs.pnct13.n.log 2.708050
## D.wrds.stop.n.log 2.564949
## min.sold.fctr.predict.Final.glmnet.N
## D.chrs.n.log 0
## D.chrs.pnct13.n.log 0
## D.wrds.stop.n.log 0
## min.sold.fctr.predict.Final.glmnet.Y
## D.chrs.n.log 0
## D.chrs.pnct13.n.log 0
## D.wrds.stop.n.log 0
## [1] "newobs sold.fctr.predict.Final.glmnet N: min < min of Train range: 1"
## UniqueID sold.fctr.predict.Final.glmnet sprice.d20nexp
## 2614 12625 N 1.929714e-22
## id cor.y exclude.as.feat cor.y.abs
## sprice.d20nexp sprice.d20nexp 0.3979951 FALSE 0.3979951
## cor.high.X freqRatio percentUnique zeroVar nzv
## sprice.d20nexp sprice.log10 2.807692 30.21622 FALSE FALSE
## is.cor.y.abs.low interaction.feat shapiro.test.p.value
## sprice.d20nexp FALSE <NA> 1.494597e-59
## rsp_var_raw id_var rsp_var max min
## sprice.d20nexp FALSE NA NA 0.9995001 1.929714e-22
## max.sold.fctr.N max.sold.fctr.Y min.sold.fctr.N
## sprice.d20nexp 0.9517052 0.9995001 2.027639e-22
## min.sold.fctr.Y max.sold.fctr.predict.All.X.glmnet.N
## sprice.d20nexp 2.200702e-15 0.8611384
## max.sold.fctr.predict.All.X.glmnet.Y
## sprice.d20nexp 0.9995001
## min.sold.fctr.predict.All.X.glmnet.N
## sprice.d20nexp 9.484101e-19
## min.sold.fctr.predict.All.X.glmnet.Y
## sprice.d20nexp 2.087968e-14
## max.sold.fctr.predict.Final.glmnet.N
## sprice.d20nexp 0.7538966
## max.sold.fctr.predict.Final.glmnet.Y
## sprice.d20nexp 0.9995001
## min.sold.fctr.predict.Final.glmnet.N
## sprice.d20nexp 1.929714e-22
## min.sold.fctr.predict.Final.glmnet.Y
## sprice.d20nexp 2.681004e-14
## [1] "newobs sold.fctr.predict.Final.glmnet N: max > max of Train range: 1"
## UniqueID sold.fctr.predict.Final.glmnet D.wrds.stop.n.log
## 2614 12625 N 2.397895
## sprice.log10 sprice.root2
## 2614 6.907745 31.62262
## id cor.y exclude.as.feat
## D.wrds.stop.n.log D.wrds.stop.n.log 0.003851183 FALSE
## sprice.log10 sprice.log10 -0.469398937 FALSE
## sprice.root2 sprice.root2 -0.511275385 FALSE
## cor.y.abs cor.high.X freqRatio percentUnique zeroVar
## D.wrds.stop.n.log 0.003851183 <NA> 8.732484 0.5405405 FALSE
## sprice.log10 0.469398937 sprice.root2 2.807692 30.2162162 FALSE
## sprice.root2 0.511275385 <NA> 2.807692 30.2162162 FALSE
## nzv is.cor.y.abs.low interaction.feat
## D.wrds.stop.n.log FALSE FALSE <NA>
## sprice.log10 FALSE FALSE <NA>
## sprice.root2 FALSE FALSE <NA>
## shapiro.test.p.value rsp_var_raw id_var rsp_var
## D.wrds.stop.n.log 4.444633e-54 FALSE NA NA
## sprice.log10 1.697609e-47 FALSE NA NA
## sprice.root2 6.339275e-16 FALSE NA NA
## max min max.sold.fctr.N max.sold.fctr.Y
## D.wrds.stop.n.log 2.564949 0.00000 2.302585 2.302585
## sprice.log10 6.907745 -4.60517 6.906755 6.514713
## sprice.root2 31.622618 0.10000 31.606961 25.980762
## min.sold.fctr.N min.sold.fctr.Y
## D.wrds.stop.n.log 0.00000000 0.00000
## sprice.log10 -0.01005034 -4.60517
## sprice.root2 0.99498744 0.10000
## max.sold.fctr.predict.All.X.glmnet.N
## D.wrds.stop.n.log 2.302585
## sprice.log10 6.721414
## sprice.root2 28.809547
## max.sold.fctr.predict.All.X.glmnet.Y
## D.wrds.stop.n.log 2.302585
## sprice.log10 6.445720
## sprice.root2 25.099801
## min.sold.fctr.predict.All.X.glmnet.N
## D.wrds.stop.n.log 0.000000
## sprice.log10 1.095273
## sprice.root2 1.729162
## min.sold.fctr.predict.All.X.glmnet.Y
## D.wrds.stop.n.log 0.00000
## sprice.log10 -4.60517
## sprice.root2 0.10000
## max.sold.fctr.predict.Final.glmnet.N
## D.wrds.stop.n.log 2.397895
## sprice.log10 6.907745
## sprice.root2 31.622618
## max.sold.fctr.predict.Final.glmnet.Y
## D.wrds.stop.n.log 2.564949
## sprice.log10 6.437752
## sprice.root2 25.000000
## min.sold.fctr.predict.Final.glmnet.N
## D.wrds.stop.n.log 0.000000
## sprice.log10 1.731656
## sprice.root2 2.376973
## min.sold.fctr.predict.Final.glmnet.Y
## D.wrds.stop.n.log 0.00000
## sprice.log10 -4.60517
## sprice.root2 0.10000
## [1] "newobs total range outliers: 5"#stop(here"); glb_to_sav(); sav_obsout_df <- obsout_df; all.equal(sav_obsout_df, obsout_df); obsout_df <- sav_obsout_df
if (!is.null(glbOutStackFnames)) {
for (fname in glbOutStackFnames) {
print(sprintf("Stacking file %s to prediction output...", fname))
#obsout_df <- dplyr::arrange_(rbind(obsout_df, read.csv(fname)), "UniqueID")
obsout_df <- dplyr::arrange_(rbind(obsout_df,
#read.csv(fname) %>% filter(!(UniqueID %in% obsout_df$UniqueID))),
#read.csv(fname) %>% filter(!(UniqueID %in% obsout_df[, glb_id_var]))),
read.csv(fname) %>%
dplyr::filter_(interp(~!(var %in% obsout_df$var),
var = as.name(glb_id_var)))),
glb_id_var)
if (nrow(obsout_df) != length(unique(obsout_df[, glb_id_var])))
stop("Potential dups in stacked prediction output")
}
}
out_fname <- paste0(glb_out_pfx, "out.csv")
write.csv(obsout_df, out_fname, quote = FALSE, row.names = FALSE)
#cat(" ", "\n", file=submit_fn, append=TRUE)
# print(orderBy(~ -max.auc.OOB, glb_models_df[, c("id",
# "max.auc.OOB", "max.Accuracy.OOB")]))
for (txt_var in glbFeatsText) {
# Print post-stem-words but need post-stop-words for debugging ?
print(sprintf(" All post-stem-words term weights for %s:", txt_var))
myprint_df(glb_post_stem_words_terms_df_lst[[txt_var]])
terms_mtrx <- glb_post_stem_words_terms_mtrx_lst[[txt_var]]
print(glb_allobs_df[
which(terms_mtrx[, tail(glb_post_stem_words_terms_df_lst[[txt_var]], 1)$pos] > 0),
c(glb_id_var, glbFeatsText)])
print(nrow(subset(glb_post_stem_words_terms_df_lst[[txt_var]], freq == 1)))
#print(glb_allobs_df[which(terms_mtrx[, 207] > 0), c(glb_id_var, glbFeatsText)])
#unlist(strsplit(glb_allobs_df[2157, "description"], ""))
#glb_allobs_df[2442, c(glb_id_var, glbFeatsText)]
#terms_mtrx[2442, terms_mtrx[2442, ] > 0]
print(sprintf(" All post-stem-words term freq distribution for %s:", txt_var))
print(table(glb_post_stem_words_terms_df_lst[[txt_var]]$freq))
print(sprintf(" All post-stem-words term length distribution for %s:", txt_var))
print(table(nchar(glb_post_stem_words_terms_df_lst[[txt_var]]$term)))
print(subset(glb_post_stem_words_terms_df_lst[[txt_var]], nchar(term) >= 10))
print(sprintf(" Analyzed term weights for %s:", txt_var))
tmp_df <- glb_post_stem_words_terms_df_lst[[txt_var]]
anl_terms_vctr <- union(select_terms, assoc_terms)
print(subset(tmp_df, term %in% anl_terms_vctr))
# tmp_freq1_df <- subset(tmp_df, freq == 1)
# tmp_freq1_df$top_n <- grepl(paste0(top_n_vctr, collapse="|"), tmp_freq1_df$term)
# print(subset(tmp_freq1_df, top_n == TRUE))
}## [1] " All post-stem-words term weights for descr.my:"
## term weight freq pos cor.y cor.y.abs chisq.stat
## condit condit 160.55353 492 52 -0.033248924 0.033248924 68.76178
## likenew likenew 123.78010 70 117 -0.043168766 0.043168766 22.07653
## use use 111.06226 285 210 0.003025811 0.003025811 45.28121
## in in 102.43546 432 98 -0.071653919 0.071653919 40.30764
## and and 92.37461 331 20 0.006010357 0.006010357 38.45444
## is is 87.72451 335 110 -0.042931491 0.042931491 26.34374
## chisq.pval nzv.freqRatio nzv.percentUnique nzv weight.N
## condit 3.361980e-06 36.82927 1.351351 TRUE 61.51671
## likenew 1.817939e-01 257.71429 0.972973 TRUE 54.11495
## use 1.518411e-02 86.63158 1.513514 TRUE 40.77891
## in 1.981535e-02 33.52174 1.351351 TRUE 46.12932
## and 7.095553e-02 55.24138 1.513514 TRUE 35.73653
## is 5.541564e-01 54.06667 1.567568 TRUE 35.26992
## weight.Y weight.NA
## condit 43.14002 55.89681
## likenew 19.52427 50.14088
## use 35.93377 34.34958
## in 26.82731 29.47884
## and 31.78712 24.85096
## is 23.45595 28.99864
## term weight freq pos cor.y cor.y.abs chisq.stat
## has has 59.117295 176 91 0.041309583 0.041309583 19.172726
## day day 16.506195 31 59 -0.031464000 0.031464000 10.200442
## owner owner 11.215175 14 151 -0.010637108 0.010637108 5.718283
## ipadair2 ipadair2 8.168607 9 106 0.017806052 0.017806052 4.057710
## can can 7.358665 12 43 0.009697873 0.009697873 6.069230
## wifion wifion 4.580205 6 221 -0.035036304 0.035036304 2.582077
## chisq.pval nzv.freqRatio nzv.percentUnique nzv weight.N
## has 0.5740608 86.300 1.1891892 TRUE 18.221776
## day 0.2512385 228.125 0.4864865 TRUE 8.657983
## owner 0.2211980 614.000 0.2702703 TRUE 4.254375
## ipadair2 0.5411374 922.000 0.3243243 TRUE 2.196428
## can 0.4154798 1844.000 0.3783784 TRUE 1.550491
## wifion 0.4606404 1847.000 0.2162162 TRUE 2.164131
## weight.Y weight.NA
## has 21.925360 18.970159
## day 4.166689 3.681523
## owner 2.417015 4.543785
## ipadair2 3.820810 2.151369
## can 1.885367 3.922808
## wifion 0.000000 2.416074
## term weight freq pos cor.y cor.y.abs chisq.stat
## 4g 4g 2.200200 3 9 0.0002965679 0.0002965679 2.023052e+00
## sprint sprint 2.037099 2 191 -0.0215577316 0.0215577316 1.281078e-27
## verizon verizon 2.037099 2 212 -0.0215577316 0.0215577316 1.281078e-27
## tmobil tmobil 1.893062 2 203 0.0094480109 0.0094480109 2.023052e+00
## ipad1 ipad1 1.834814 2 101 0.0067428569 0.0067428569 2.023052e+00
## gold gold 1.137069 1 86 -0.0215577316 0.0215577316 1.281078e-27
## chisq.pval nzv.freqRatio nzv.percentUnique nzv weight.N
## 4g 0.3636637 1848 0.1621622 TRUE 0.6989804
## sprint 1.0000000 1849 0.1081081 TRUE 0.7407634
## verizon 1.0000000 1849 0.1081081 TRUE 0.7407634
## tmobil 0.3636637 1848 0.1621622 TRUE 0.7407634
## ipad1 0.3636637 1848 0.1621622 TRUE 0.7977452
## gold 1.0000000 1849 0.1081081 TRUE 1.1370687
## weight.Y weight.NA
## 4g 0.6116078 0.8896114
## sprint 0.0000000 1.2963359
## verizon 0.0000000 1.2963359
## tmobil 1.1522986 0.0000000
## ipad1 1.0370687 0.0000000
## gold 0.0000000 0.0000000
## UniqueID
## 1177 11183
## descr.my
## 1177 *** Likenew, Apple iPadAir2 Gold 128GB. Comes with AC adapter and lightning cable.
## [1] 1
## [1] " All post-stem-words term freq distribution for descr.my:"
##
## 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19
## 1 4 3 3 1 2 1 1 13 8 13 6 13 5 9 5 6 8
## 20 21 22 23 24 25 26 27 28 30 31 32 34 35 36 37 38 40
## 2 5 1 3 1 2 5 8 3 1 7 2 3 1 1 3 2 2
## 42 43 44 45 46 50 51 52 53 54 55 57 58 59 61 63 64 65
## 1 2 1 2 2 1 1 2 1 1 1 2 3 1 1 1 1 1
## 67 70 72 75 77 78 79 80 81 82 88 97 99 100 101 106 109 117
## 1 1 2 2 3 1 1 1 1 2 1 1 1 1 1 1 2 2
## 125 129 151 154 160 167 176 197 198 208 211 214 249 279 285 286 331 335
## 2 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1
## 432 492
## 1 1
## [1] " All post-stem-words term length distribution for descr.my:"
##
## 1 2 3 4 5 6 7 8 9 10
## 6 17 28 64 51 29 15 8 6 2
## term weight freq pos cor.y cor.y.abs chisq.stat
## profession profession 9.495620 16 160 -0.07142724 0.07142724 9.508800
## manufactur manufactur 8.616795 8 123 0.01332077 0.01332077 6.928535
## chisq.pval nzv.freqRatio nzv.percentUnique nzv weight.N
## profession 0.04956674 613.0 0.2702703 TRUE 6.606743
## manufactur 0.32750691 921.5 0.3783784 TRUE 3.089182
## weight.Y weight.NA
## profession 0.000000 2.888877
## manufactur 4.481277 1.046336
## [1] " Analyzed term weights for descr.my:"
## term weight freq pos cor.y cor.y.abs chisq.stat
## condit condit 160.55353 492 52 -0.033248924 0.033248924 68.76178
## use use 111.06226 285 210 0.003025811 0.003025811 45.28121
## in in 102.43546 432 98 -0.071653919 0.071653919 40.30764
## good good 86.95851 197 87 0.016394675 0.016394675 33.86257
## screen screen 73.07057 208 172 0.040605274 0.040605274 32.79902
## with with 63.69164 214 223 -0.066087601 0.066087601 33.25670
## of of 52.99927 151 139 0.052259530 0.052259530 34.49697
## mint mint 52.07871 63 128 -0.055691207 0.055691207 32.30119
## or or 51.11272 125 144 -0.022614420 0.022614420 36.49790
## cosmet cosmet 43.18436 117 55 -0.088480863 0.088480863 32.40489
## minor minor 41.27081 117 127 -0.042692360 0.042692360 33.12974
## light light 36.41889 81 116 -0.032814746 0.032814746 35.12786
## 100 100 27.52852 64 3 -0.111398502 0.111398502 30.45739
## from from 24.73125 57 78 -0.030653359 0.030653359 34.28905
## hous hous 24.59165 72 94 -0.129274878 0.129274878 41.95872
## chisq.pval nzv.freqRatio nzv.percentUnique nzv weight.N
## condit 3.361980e-06 36.82927 1.3513514 TRUE 61.51671
## use 1.518411e-02 86.63158 1.5135135 TRUE 40.77891
## in 1.981535e-02 33.52174 1.3513514 TRUE 46.12932
## good 3.748545e-02 95.16667 1.1891892 TRUE 29.35392
## screen 3.549380e-02 70.91667 1.1351351 TRUE 23.12678
## with 1.553541e-02 46.88889 1.0270270 TRUE 30.89756
## of 2.295297e-02 145.00000 1.1351351 TRUE 16.02233
## mint 2.024848e-02 94.89474 1.0270270 TRUE 24.27170
## or 9.163373e-03 87.95000 1.0810811 TRUE 22.27352
## cosmet 3.507470e-03 88.40000 0.8108108 TRUE 22.73428
## minor 1.085239e-02 92.52632 0.9729730 TRUE 20.58455
## light 3.817238e-03 99.27778 0.9189189 TRUE 17.42219
## 100 1.342479e-03 180.50000 0.6486486 TRUE 18.16396
## from 6.072414e-04 95.31579 0.7027027 TRUE 10.97996
## hous 3.344151e-06 81.68182 0.5405405 TRUE 17.25642
## weight.Y weight.NA
## condit 43.140019 55.896806
## use 35.933771 34.349578
## in 26.827307 29.478838
## good 30.064015 27.540573
## screen 28.028461 21.915331
## with 16.195788 16.598293
## of 21.710580 15.266363
## mint 6.994811 20.812208
## or 15.175342 13.663856
## cosmet 7.055599 13.394481
## minor 11.627271 9.058990
## light 10.171779 8.824924
## 100 1.314682 8.049871
## from 5.861552 7.889737
## hous 1.193859 6.141364if (glb_is_classification && glb_is_binomial)
print(glb_models_df[glb_models_df$id == glb_sel_mdl_id,
"opt.prob.threshold.OOB"])## [1] 0.3print(sprintf("glb_sel_mdl_id: %s", glb_sel_mdl_id))## [1] "glb_sel_mdl_id: All.X.glmnet"print(sprintf("glb_fin_mdl_id: %s", glb_fin_mdl_id))## [1] "glb_fin_mdl_id: Final.glmnet"get_dsp_models_df()## [1] "Cross Validation issues:"## Warning in get_dsp_models_df(): Cross Validation issues:## MFO.myMFO_classfr Random.myrandom_classfr
## 0 0
## Max.cor.Y.rcv.1X1.glmnet Max.cor.Y.rcv.1X1.cp.0.rpart
## 0 0## id
## RFE.X.glm RFE.X.glm
## Max.cor.Y.rcv.1X1.glmnet Max.cor.Y.rcv.1X1.glmnet
## Interact.High.cor.Y.glmnet Interact.High.cor.Y.glmnet
## Max.cor.Y.rpart Max.cor.Y.rpart
## Low.cor.X.glmnet Low.cor.X.glmnet
## Max.cor.Y.rcv.3X3.glmnet Max.cor.Y.rcv.3X3.glmnet
## Max.cor.Y.rcv.1X1.Interact.glmnet Max.cor.Y.rcv.1X1.Interact.glmnet
## Max.cor.Y.rcv.3X1.glmnet Max.cor.Y.rcv.3X1.glmnet
## Max.cor.Y.rcv.3X5.glmnet Max.cor.Y.rcv.3X5.glmnet
## Max.cor.Y.rcv.5X1.glmnet Max.cor.Y.rcv.5X1.glmnet
## Max.cor.Y.rcv.5X3.glmnet Max.cor.Y.rcv.5X3.glmnet
## Max.cor.Y.rcv.5X5.glmnet Max.cor.Y.rcv.5X5.glmnet
## RFE.X.glmnet RFE.X.glmnet
## All.X.glmnet All.X.glmnet
## Max.cor.Y.rcv.1X1.cp.0.rpart Max.cor.Y.rcv.1X1.cp.0.rpart
## MFO.myMFO_classfr MFO.myMFO_classfr
## Random.myrandom_classfr Random.myrandom_classfr
## Final.glmnet Final.glmnet
## Final.glm Final.glm
## max.Accuracy.OOB max.AUCROCR.OOB
## RFE.X.glm 0.7879518 0.8700475
## Max.cor.Y.rcv.1X1.glmnet 0.7722892 0.8197236
## Interact.High.cor.Y.glmnet 0.7698795 0.8189000
## Max.cor.Y.rpart 0.7698795 0.7855854
## Low.cor.X.glmnet 0.7686747 0.8727052
## Max.cor.Y.rcv.3X3.glmnet 0.7662651 0.8197470
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.7662651 0.8197470
## Max.cor.Y.rcv.3X1.glmnet 0.7662651 0.8197353
## Max.cor.Y.rcv.3X5.glmnet 0.7650602 0.8196593
## Max.cor.Y.rcv.5X1.glmnet 0.7638554 0.8195600
## Max.cor.Y.rcv.5X3.glmnet 0.7638554 0.8195600
## Max.cor.Y.rcv.5X5.glmnet 0.7638554 0.8193790
## RFE.X.glmnet 0.7614458 0.8721444
## All.X.glmnet 0.7614458 0.8721444
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.7433735 0.7934621
## MFO.myMFO_classfr 0.4614458 0.5000000
## Random.myrandom_classfr 0.4614458 0.4956046
## Final.glmnet NA NA
## Final.glm NA NA
## max.AUCpROC.OOB max.Accuracy.fit
## RFE.X.glm 0.7851882 0.8307135
## Max.cor.Y.rcv.1X1.glmnet 0.7665376 0.7843137
## Interact.High.cor.Y.glmnet 0.7551533 0.8026098
## Max.cor.Y.rpart 0.7445254 0.7964073
## Low.cor.X.glmnet 0.7801532 0.8349657
## Max.cor.Y.rcv.3X3.glmnet 0.7603840 0.7954221
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.7603840 0.7954221
## Max.cor.Y.rcv.3X1.glmnet 0.7592654 0.7950911
## Max.cor.Y.rcv.3X5.glmnet 0.7603840 0.7962655
## Max.cor.Y.rcv.5X1.glmnet 0.7615026 0.7990899
## Max.cor.Y.rcv.5X3.glmnet 0.7615026 0.7964364
## Max.cor.Y.rcv.5X5.glmnet 0.7641135 0.7970687
## RFE.X.glmnet 0.7835089 0.8303886
## All.X.glmnet 0.7835089 0.8303886
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.7504950 0.8323529
## MFO.myMFO_classfr 0.5000000 0.4627451
## Random.myrandom_classfr 0.5064252 0.4627451
## Final.glmnet NA 0.8203656
## Final.glm NA 0.8185584
## opt.prob.threshold.fit
## RFE.X.glm 0.4
## Max.cor.Y.rcv.1X1.glmnet 0.4
## Interact.High.cor.Y.glmnet 0.4
## Max.cor.Y.rpart 0.5
## Low.cor.X.glmnet 0.5
## Max.cor.Y.rcv.3X3.glmnet 0.5
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.5
## Max.cor.Y.rcv.3X1.glmnet 0.5
## Max.cor.Y.rcv.3X5.glmnet 0.5
## Max.cor.Y.rcv.5X1.glmnet 0.5
## Max.cor.Y.rcv.5X3.glmnet 0.5
## Max.cor.Y.rcv.5X5.glmnet 0.5
## RFE.X.glmnet 0.4
## All.X.glmnet 0.4
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.4
## MFO.myMFO_classfr 0.4
## Random.myrandom_classfr 0.4
## Final.glmnet 0.4
## Final.glm 0.4
## opt.prob.threshold.OOB
## RFE.X.glm 0.5
## Max.cor.Y.rcv.1X1.glmnet 0.4
## Interact.High.cor.Y.glmnet 0.3
## Max.cor.Y.rpart 0.3
## Low.cor.X.glmnet 0.3
## Max.cor.Y.rcv.3X3.glmnet 0.4
## Max.cor.Y.rcv.1X1.Interact.glmnet 0.4
## Max.cor.Y.rcv.3X1.glmnet 0.4
## Max.cor.Y.rcv.3X5.glmnet 0.4
## Max.cor.Y.rcv.5X1.glmnet 0.4
## Max.cor.Y.rcv.5X3.glmnet 0.4
## Max.cor.Y.rcv.5X5.glmnet 0.4
## RFE.X.glmnet 0.3
## All.X.glmnet 0.3
## Max.cor.Y.rcv.1X1.cp.0.rpart 0.3
## MFO.myMFO_classfr 0.4
## Random.myrandom_classfr 0.4
## Final.glmnet NA
## Final.glm NAif (glb_is_regression) {
print(sprintf("%s OOB RMSE: %0.4f", glb_sel_mdl_id,
glb_models_df[glb_models_df$id == glb_sel_mdl_id, "min.RMSE.OOB"]))
if (!is.null(glb_category_var)) {
#stop(here"); glb_to_sav(); glb_ctgry_df <- sav_ctgry_df
# OOB_ctgry_df <- myget_category_stats(glb_OOBobs_df, glb_sel_mdl_id, "OOB")
# glb_ctgry_df <- merge(glb_ctgry_df, subset(OOB_ctgry_df, select=-.n.OOB),
# by=glb_category_var, all=TRUE)
#
# #glb_fitobs_df <- glb_get_predictions(glb_fitobs_df, glb_sel_mdl_id, glb_rsp_var_out)
# glb_ctgry_df <- merge(glb_ctgry_df,
# myget_category_stats(obs_df=glb_fitobs_df, mdl_id=glb_sel_mdl_id, label="fit"),
# by=glb_category_var, all=TRUE)
# row.names(glb_ctgry_df) <- glb_ctgry_df[, glb_category_var]
glb_ctgry_df <- merge(glb_ctgry_df,
myget_category_stats(obs_df=glb_trnobs_df, mdl_id=glb_fin_mdl_id, label="trn"),
by=glb_category_var, all=TRUE)
row.names(glb_ctgry_df) <- glb_ctgry_df[, glb_category_var]
glb_ctgry_df <- merge(glb_ctgry_df,
myget_category_stats(obs_df=glb_newobs_df, mdl_id=glb_fin_mdl_id, label="new"),
by=glb_category_var, all=TRUE)
row.names(glb_ctgry_df) <- glb_ctgry_df[, glb_category_var]
if (any(grepl("OOB", glbMdlMetricsEval)))
print(orderBy(~-err.abs.OOB.mean, glb_ctgry_df)) else
print(orderBy(~-err.abs.fit.mean, glb_ctgry_df))
print(colSums(glb_ctgry_df[, -grep(glb_category_var, names(glb_ctgry_df))]))
}
if ((glb_rsp_var %in% names(glb_newobs_df)) &&
!(any(is.na(glb_newobs_df[, glb_rsp_var])))) {
pred_stats_df <-
mypredict_mdl(mdl=glb_models_lst[[glb_fin_mdl_id]],
df=glb_newobs_df,
rsp_var=glb_rsp_var,
rsp_var_out=glb_rsp_var_out,
mdl_id=glb_fin_mdl_id,
label="new",
model_summaryFunction=glb_sel_mdl$control$summaryFunction,
model_metric=glb_sel_mdl$metric,
model_metric_maximize=glb_sel_mdl$maximize,
ret_type="stats")
print(sprintf("%s prediction stats for glb_newobs_df:", glb_fin_mdl_id))
print(pred_stats_df)
}
}
if (glb_is_classification) {
print(sprintf("%s OOB confusion matrix & accuracy: ", glb_sel_mdl_id))
print(t(confusionMatrix(glb_OOBobs_df[, paste0(glb_rsp_var_out, glb_sel_mdl_id)],
glb_OOBobs_df[, glb_rsp_var])$table))
if (!is.null(glb_category_var)) {
glb_ctgry_df <- merge(glb_ctgry_df,
myget_category_stats(obs_df=glb_trnobs_df, mdl_id=glb_fin_mdl_id, label="trn"),
by=glb_category_var, all=TRUE)
row.names(glb_ctgry_df) <- glb_ctgry_df[, glb_category_var]
glb_ctgry_df <- merge(glb_ctgry_df,
myget_category_stats(obs_df=glb_newobs_df, mdl_id=glb_fin_mdl_id, label="new"),
by=glb_category_var, all=TRUE)
row.names(glb_ctgry_df) <- glb_ctgry_df[, glb_category_var]
if (any(grepl("OOB", glbMdlMetricsEval)))
print(orderBy(~-err.abs.OOB.mean, glb_ctgry_df[, -1])) else
print(orderBy(~-err.abs.fit.mean, glb_ctgry_df[, -1]))
print(colSums(glb_ctgry_df[, -grep(glb_category_var, names(glb_ctgry_df))]))
# print("Top category OOB errors:")
# print(glb_OOBobs_df[(glb_OOBobs_df[, glb_category_var] ==
# dsp_ctgry_df[1, glb_category_var]) &
# !glb_OOBobs_df[, predct_accurate_var_name],
# c(glb_id_var, glb_rsp_var_raw, paste0(glb_rsp_var_out, glb_sel_mdl_id),
# glb_category_var,
# row.names(head(myget_feats_importance(glb_sel_mdl), 5)),
# # "biddable", "startprice", "condition",
# glbFeatsText)])
}
if ((glb_rsp_var %in% names(glb_newobs_df)) &&
!(any(is.na(glb_newobs_df[, glb_rsp_var])))) {
print(sprintf("%s new confusion matrix & accuracy: ", glb_fin_mdl_id))
print(t(confusionMatrix(glb_newobs_df[, paste0(glb_rsp_var_out, glb_fin_mdl_id)],
glb_newobs_df[, glb_rsp_var])$table))
}
} ## [1] "All.X.glmnet OOB confusion matrix & accuracy: "
## Prediction
## Reference N Y
## N 309 138
## Y 60 323
## .n.OOB .n.Fit .n.Tst .freqRatio.Fit .freqRatio.OOB
## iPad1 95 130 88 0.12745098 0.11445783
## iPadmini2 53 54 56 0.05294118 0.06385542
## Unknown 96 108 92 0.10588235 0.11566265
## iPadmini 123 154 111 0.15098039 0.14819277
## iPadAir2 69 102 62 0.10000000 0.08313253
## iPad2 142 144 154 0.14117647 0.17108434
## iPad3 61 92 55 0.09019608 0.07349398
## iPadmini3 36 54 38 0.05294118 0.04337349
## iPadAir 82 98 74 0.09607843 0.09879518
## iPad4 73 84 68 0.08235294 0.08795181
## .freqRatio.Tst err.abs.fit.sum err.abs.fit.mean .n.fit
## iPad1 0.11027569 23.91786 0.1839835 130
## iPadmini2 0.07017544 14.19476 0.2628659 54
## Unknown 0.11528822 26.52091 0.2455640 108
## iPadmini 0.13909774 30.21777 0.1962193 154
## iPadAir2 0.07769424 21.24831 0.2083168 102
## iPad2 0.19298246 31.86069 0.2212548 144
## iPad3 0.06892231 21.85916 0.2375996 92
## iPadmini3 0.04761905 10.64581 0.1971447 54
## iPadAir 0.09273183 13.80524 0.1408698 98
## iPad4 0.08521303 10.53643 0.1254337 84
## err.abs.OOB.sum err.abs.OOB.mean err.abs.trn.sum
## iPad1 31.738295 0.3340873 59.73383
## iPadmini2 15.149743 0.2858442 32.44337
## Unknown 25.545260 0.2660965 55.98075
## iPadmini 32.704594 0.2658910 75.52109
## iPadAir2 17.785002 0.2577536 44.49984
## iPad2 35.143654 0.2474905 74.79123
## iPad3 14.689955 0.2408189 39.63516
## iPadmini3 8.566058 0.2379461 21.58114
## iPadAir 15.249371 0.1859679 36.67291
## iPad4 13.177556 0.1805145 33.00200
## err.abs.trn.mean .n.trn err.abs.new.sum err.abs.new.mean .n.new
## iPad1 0.2654837 225 NA NA 88
## iPadmini2 0.3032091 107 NA NA 56
## Unknown 0.2744154 204 NA NA 92
## iPadmini 0.2726393 277 NA NA 111
## iPadAir2 0.2602330 171 NA NA 62
## iPad2 0.2615078 286 NA NA 154
## iPad3 0.2590533 153 NA NA 55
## iPadmini3 0.2397904 90 NA NA 38
## iPadAir 0.2037384 180 NA NA 74
## iPad4 0.2102038 157 NA NA 68
## .n.OOB .n.Fit .n.Tst .freqRatio.Fit
## 830.000000 1020.000000 798.000000 1.000000
## .freqRatio.OOB .freqRatio.Tst err.abs.fit.sum err.abs.fit.mean
## 1.000000 1.000000 204.806955 2.019252
## .n.fit err.abs.OOB.sum err.abs.OOB.mean err.abs.trn.sum
## 1020.000000 209.749487 2.502411 473.861322
## err.abs.trn.mean .n.trn err.abs.new.sum err.abs.new.mean
## 2.550274 1850.000000 NA NA
## .n.new
## 798.000000if (!is.null(glb_featsimp_df))
print(orderBy(as.formula(paste0("~ -", glb_sel_mdl_id, ".importance")),
subset(glb_featsimp_df, importance > 10)))## All.X.glmnet.importance
## spdiff.cut.fctr(-1,0] 100.00000
## spdiff.cut.fctr(1,10] 91.77825
## spdiff.cut.fctr(0,1] 91.77769
## spdiff.cut.fctr(-10,-1] 89.53891
## spdiff.cut.fctr(10,100] 86.86797
## cellular.fctr1:carrier.fctrOther 75.82223
## biddable 75.17563
## spdiff.cut.fctr(-100,-10] 73.61063
## prdl.my.fctriPadAir2 69.61045
## storage.fctrUnknown 68.07849
## prdl.my.fctriPadAir 66.43833
## prdl.my.fctriPad4 65.16814
## prdl.my.fctriPadmini2 64.52272
## D.wrds.n.log 63.88632
## spdiff.cut.fctr(100,1e+03] 62.38931
## cellular.fctr1:carrier.fctrSprint 61.81532
## prdl.my.fctrUnknown:.clusterid.fctr2 60.62908
## prdl.my.fctriPadmini3 59.42417
## condition.fctrNew other (see details) 59.12005
## D.chrs.pnct11.n.log 56.89033
## color.fctrUnknown 55.86778
## cellular.fctr1:carrier.fctrVerizon 55.50911
## prdl.my.fctriPad3:.clusterid.fctr2 55.10589
## storage.fctr64 54.64454
## prdl.my.fctriPad3 54.34045
## cellular.fctr0:carrier.fctrNone 54.22434
## startprice.dgt2.is9 53.73722
## D.wrds.stop.n.log 53.73265
## color.fctrGold 53.48381
## condition.fctrNew 52.97506
## startprice.dgt1.is9 52.93673
## storage.fctr32 52.71638
## cellular.fctr1 52.71142
## D.weight.post.stop.sum 52.71142
## cellular.fctr0:carrier.fctrOther 52.71142
## cellular.fctr0:carrier.fctrSprint 52.71142
## cellular.fctr0:carrier.fctrT-Mobile 52.71142
## cellular.fctr0:carrier.fctrUnknown 52.71142
## cellular.fctr0:carrier.fctrVerizon 52.71142
## cellular.fctr1:carrier.fctrNone 52.71142
## cellular.fctrUnknown:carrier.fctrNone 52.71142
## cellular.fctrUnknown:carrier.fctrOther 52.71142
## cellular.fctrUnknown:carrier.fctrSprint 52.71142
## cellular.fctrUnknown:carrier.fctrT-Mobile 52.71142
## cellular.fctrUnknown:carrier.fctrVerizon 52.71142
## prdl.my.fctriPad2:.clusterid.fctr2 52.71142
## prdl.my.fctriPad2:.clusterid.fctr3 52.71142
## prdl.my.fctriPad3:.clusterid.fctr3 52.71142
## prdl.my.fctriPadAir2:.clusterid.fctr2 52.71142
## prdl.my.fctriPadAir2:.clusterid.fctr3 52.71142
## prdl.my.fctriPadAir:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini2:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini3:.clusterid.fctr2 52.71142
## prdl.my.fctriPadmini3:.clusterid.fctr3 52.71142
## prdl.my.fctriPadmini:.clusterid.fctr2 52.71142
## prdl.my.fctriPadmini:.clusterid.fctr3 52.71142
## D.chrs.uppr.n.log 52.60858
## cellular.fctr1:carrier.fctrUnknown 52.55260
## .rnorm 52.38568
## D.ratio.weight.sum.wrds.n 52.18561
## startprice.dcm2.is9 51.96783
## sprice.root2 51.74950
## D.weight.sum 51.66462
## D.weight.post.stem.sum 51.63919
## storage.fctr16 51.55882
## D.chrs.pnct13.n.log 51.24550
## D.wrds.unq.n.log 51.14189
## color.fctrWhite 50.71378
## cellular.fctr1:carrier.fctrT-Mobile 50.21020
## D.chrs.n.log 50.18255
## D.terms.post.stem.n.log 50.17209
## prdl.my.fctriPadAir:.clusterid.fctr2 49.80426
## color.fctrSpace Gray 49.63204
## prdl.my.fctriPad2 48.61429
## prdl.my.fctriPadmini 48.49638
## startprice.dcm1.is9 48.37251
## sprice.log10 47.59704
## prdl.my.fctriPadmini2:.clusterid.fctr2 47.13793
## prdl.my.fctriPad1 46.63375
## condition.fctrFor parts or not working 46.18220
## prdl.my.fctriPad1:.clusterid.fctr2 46.14804
## prdl.my.fctrUnknown:.clusterid.fctr3 45.97618
## cellular.fctrUnknown:carrier.fctrUnknown 45.72360
## cellular.fctrUnknown 45.68888
## D.terms.post.stop.n.log 44.49862
## condition.fctrSeller refurbished 43.60662
## prdl.my.fctriPad1:.clusterid.fctr3 42.66158
## condition.fctrManufacturer refurbished 42.10131
## prdl.my.fctriPad4:.clusterid.fctr2 39.81232
## sprice.d20nexp 38.83227
## D.weight.sum.stem.stop.Ratio 26.99368
## D.ratio.wrds.stop.n.wrds.n 26.94246
## importance
## spdiff.cut.fctr(-1,0] 100.00000
## spdiff.cut.fctr(1,10] 80.40773
## spdiff.cut.fctr(0,1] 85.31015
## spdiff.cut.fctr(-10,-1] 84.15852
## spdiff.cut.fctr(10,100] 72.75819
## cellular.fctr1:carrier.fctrOther 38.39431
## biddable 71.47982
## spdiff.cut.fctr(-100,-10] 57.67475
## prdl.my.fctriPadAir2 53.70014
## storage.fctrUnknown 38.39431
## prdl.my.fctriPadAir 48.86583
## prdl.my.fctriPad4 41.79089
## prdl.my.fctriPadmini2 41.25744
## D.wrds.n.log 38.39431
## spdiff.cut.fctr(100,1e+03] 42.57950
## cellular.fctr1:carrier.fctrSprint 41.47670
## prdl.my.fctrUnknown:.clusterid.fctr2 45.06580
## prdl.my.fctriPadmini3 38.39431
## condition.fctrNew other (see details) 45.89626
## D.chrs.pnct11.n.log 38.38991
## color.fctrUnknown 38.39431
## cellular.fctr1:carrier.fctrVerizon 38.39431
## prdl.my.fctriPad3:.clusterid.fctr2 34.83380
## storage.fctr64 42.47779
## prdl.my.fctriPad3 38.39431
## cellular.fctr0:carrier.fctrNone 38.39431
## startprice.dgt2.is9 38.39431
## D.wrds.stop.n.log 38.39431
## color.fctrGold 38.39431
## condition.fctrNew 38.39431
## startprice.dgt1.is9 38.39431
## storage.fctr32 38.33297
## cellular.fctr1 39.59945
## D.weight.post.stop.sum 38.39431
## cellular.fctr0:carrier.fctrOther 38.39431
## cellular.fctr0:carrier.fctrSprint 38.39431
## cellular.fctr0:carrier.fctrT-Mobile 38.39431
## cellular.fctr0:carrier.fctrUnknown 38.39431
## cellular.fctr0:carrier.fctrVerizon 38.39431
## cellular.fctr1:carrier.fctrNone 38.39431
## cellular.fctrUnknown:carrier.fctrNone 38.39431
## cellular.fctrUnknown:carrier.fctrOther 38.39431
## cellular.fctrUnknown:carrier.fctrSprint 38.39431
## cellular.fctrUnknown:carrier.fctrT-Mobile 38.39431
## cellular.fctrUnknown:carrier.fctrVerizon 38.39431
## prdl.my.fctriPad2:.clusterid.fctr2 38.39431
## prdl.my.fctriPad2:.clusterid.fctr3 38.39431
## prdl.my.fctriPad3:.clusterid.fctr3 38.39431
## prdl.my.fctriPadAir2:.clusterid.fctr2 38.39431
## prdl.my.fctriPadAir2:.clusterid.fctr3 38.39431
## prdl.my.fctriPadAir:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini2:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini3:.clusterid.fctr2 38.39431
## prdl.my.fctriPadmini3:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini:.clusterid.fctr2 38.39431
## prdl.my.fctriPadmini:.clusterid.fctr3 38.39431
## D.chrs.uppr.n.log 38.39431
## cellular.fctr1:carrier.fctrUnknown 38.39431
## .rnorm 38.39431
## D.ratio.weight.sum.wrds.n 38.38146
## startprice.dcm2.is9 38.39431
## sprice.root2 36.91983
## D.weight.sum 38.39431
## D.weight.post.stem.sum 38.39431
## storage.fctr16 38.39431
## D.chrs.pnct13.n.log 38.11481
## D.wrds.unq.n.log 38.39431
## color.fctrWhite 37.93000
## cellular.fctr1:carrier.fctrT-Mobile 39.35301
## D.chrs.n.log 38.39431
## D.terms.post.stem.n.log 38.39431
## prdl.my.fctriPadAir:.clusterid.fctr2 38.39431
## color.fctrSpace Gray 38.39431
## prdl.my.fctriPad2 32.84184
## prdl.my.fctriPadmini 35.61259
## startprice.dcm1.is9 33.50057
## sprice.log10 35.80737
## prdl.my.fctriPadmini2:.clusterid.fctr2 38.39431
## prdl.my.fctriPad1 30.61657
## condition.fctrFor parts or not working 34.63052
## prdl.my.fctriPad1:.clusterid.fctr2 31.35476
## prdl.my.fctrUnknown:.clusterid.fctr3 38.39379
## cellular.fctrUnknown:carrier.fctrUnknown 38.25950
## cellular.fctrUnknown 36.13550
## D.terms.post.stop.n.log 38.24626
## condition.fctrSeller refurbished 31.20851
## prdl.my.fctriPad1:.clusterid.fctr3 25.94724
## condition.fctrManufacturer refurbished 38.39431
## prdl.my.fctriPad4:.clusterid.fctr2 38.39431
## sprice.d20nexp 38.39431
## D.weight.sum.stem.stop.Ratio 38.26292
## D.ratio.wrds.stop.n.wrds.n 38.39431
## Final.glmnet.importance
## spdiff.cut.fctr(-1,0] 100.00000
## spdiff.cut.fctr(1,10] 80.40773
## spdiff.cut.fctr(0,1] 85.31015
## spdiff.cut.fctr(-10,-1] 84.15852
## spdiff.cut.fctr(10,100] 72.75819
## cellular.fctr1:carrier.fctrOther 38.39431
## biddable 71.47982
## spdiff.cut.fctr(-100,-10] 57.67475
## prdl.my.fctriPadAir2 53.70014
## storage.fctrUnknown 38.39431
## prdl.my.fctriPadAir 48.86583
## prdl.my.fctriPad4 41.79089
## prdl.my.fctriPadmini2 41.25744
## D.wrds.n.log 38.39431
## spdiff.cut.fctr(100,1e+03] 42.57950
## cellular.fctr1:carrier.fctrSprint 41.47670
## prdl.my.fctrUnknown:.clusterid.fctr2 45.06580
## prdl.my.fctriPadmini3 38.39431
## condition.fctrNew other (see details) 45.89626
## D.chrs.pnct11.n.log 38.38991
## color.fctrUnknown 38.39431
## cellular.fctr1:carrier.fctrVerizon 38.39431
## prdl.my.fctriPad3:.clusterid.fctr2 34.83380
## storage.fctr64 42.47779
## prdl.my.fctriPad3 38.39431
## cellular.fctr0:carrier.fctrNone 38.39431
## startprice.dgt2.is9 38.39431
## D.wrds.stop.n.log 38.39431
## color.fctrGold 38.39431
## condition.fctrNew 38.39431
## startprice.dgt1.is9 38.39431
## storage.fctr32 38.33297
## cellular.fctr1 39.59945
## D.weight.post.stop.sum 38.39431
## cellular.fctr0:carrier.fctrOther 38.39431
## cellular.fctr0:carrier.fctrSprint 38.39431
## cellular.fctr0:carrier.fctrT-Mobile 38.39431
## cellular.fctr0:carrier.fctrUnknown 38.39431
## cellular.fctr0:carrier.fctrVerizon 38.39431
## cellular.fctr1:carrier.fctrNone 38.39431
## cellular.fctrUnknown:carrier.fctrNone 38.39431
## cellular.fctrUnknown:carrier.fctrOther 38.39431
## cellular.fctrUnknown:carrier.fctrSprint 38.39431
## cellular.fctrUnknown:carrier.fctrT-Mobile 38.39431
## cellular.fctrUnknown:carrier.fctrVerizon 38.39431
## prdl.my.fctriPad2:.clusterid.fctr2 38.39431
## prdl.my.fctriPad2:.clusterid.fctr3 38.39431
## prdl.my.fctriPad3:.clusterid.fctr3 38.39431
## prdl.my.fctriPadAir2:.clusterid.fctr2 38.39431
## prdl.my.fctriPadAir2:.clusterid.fctr3 38.39431
## prdl.my.fctriPadAir:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini2:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini3:.clusterid.fctr2 38.39431
## prdl.my.fctriPadmini3:.clusterid.fctr3 38.39431
## prdl.my.fctriPadmini:.clusterid.fctr2 38.39431
## prdl.my.fctriPadmini:.clusterid.fctr3 38.39431
## D.chrs.uppr.n.log 38.39431
## cellular.fctr1:carrier.fctrUnknown 38.39431
## .rnorm 38.39431
## D.ratio.weight.sum.wrds.n 38.38146
## startprice.dcm2.is9 38.39431
## sprice.root2 36.91983
## D.weight.sum 38.39431
## D.weight.post.stem.sum 38.39431
## storage.fctr16 38.39431
## D.chrs.pnct13.n.log 38.11481
## D.wrds.unq.n.log 38.39431
## color.fctrWhite 37.93000
## cellular.fctr1:carrier.fctrT-Mobile 39.35301
## D.chrs.n.log 38.39431
## D.terms.post.stem.n.log 38.39431
## prdl.my.fctriPadAir:.clusterid.fctr2 38.39431
## color.fctrSpace Gray 38.39431
## prdl.my.fctriPad2 32.84184
## prdl.my.fctriPadmini 35.61259
## startprice.dcm1.is9 33.50057
## sprice.log10 35.80737
## prdl.my.fctriPadmini2:.clusterid.fctr2 38.39431
## prdl.my.fctriPad1 30.61657
## condition.fctrFor parts or not working 34.63052
## prdl.my.fctriPad1:.clusterid.fctr2 31.35476
## prdl.my.fctrUnknown:.clusterid.fctr3 38.39379
## cellular.fctrUnknown:carrier.fctrUnknown 38.25950
## cellular.fctrUnknown 36.13550
## D.terms.post.stop.n.log 38.24626
## condition.fctrSeller refurbished 31.20851
## prdl.my.fctriPad1:.clusterid.fctr3 25.94724
## condition.fctrManufacturer refurbished 38.39431
## prdl.my.fctriPad4:.clusterid.fctr2 38.39431
## sprice.d20nexp 38.39431
## D.weight.sum.stem.stop.Ratio 38.26292
## D.ratio.wrds.stop.n.wrds.n 38.39431print("glb_newobs_df prediction stats:")## [1] "glb_newobs_df prediction stats:"if (glb_is_regression)
print(myplot_histogram(glb_newobs_df, paste0(glb_rsp_var_out, glb_fin_mdl_id)))
if (glb_is_classification)
print(table(glb_newobs_df[, paste0(glb_rsp_var_out, glb_fin_mdl_id)]))##
## N Y
## 345 453# Use this to see how prediction changes by changing one or more values
# players_df <- data.frame(id=c("Chavez", "Giambi", "Menechino", "Myers", "Pena"),
# OBP=c(0.338, 0.391, 0.369, 0.313, 0.361),
# SLG=c(0.540, 0.450, 0.374, 0.447, 0.500),
# cost=c(1400000, 1065000, 295000, 800000, 300000))
# players_df$RS.predict <- predict(glb_models_lst[[csm_mdl_id]], players_df)
# print(orderBy(~ -RS.predict, players_df))
# dsp_chisq.test(Headline.contains="[Vi]deo")
if ((length(diff <- setdiff(names(glb_trnobs_df), names(glb_allobs_df))) > 0) ||
(length(diff <- setdiff(names(glb_fitobs_df), names(glb_allobs_df))) > 0) ||
(length(diff <- setdiff(names(glb_OOBobs_df), names(glb_allobs_df))) > 0) ||
(length(diff <- setdiff(names(glb_newobs_df), names(glb_allobs_df))) > 0)) {
print(diff)
stop("glb_*obs_df not in sync")
}
if (glb_save_envir)
save(glb_feats_df, glb_allobs_df,
#glb_trnobs_df, glb_fitobs_df, glb_OOBobs_df, glb_newobs_df,
glb_models_df, dsp_models_df, glb_models_lst, glb_model_type,
glb_sel_mdl, glb_sel_mdl_id,
glb_fin_mdl, glb_fin_mdl_id,
file = paste0(glb_out_pfx, "prdnew_dsk.RData"))
sav_fin_mdl <- glb_fin_mdl; sav_sel_mdl <- glb_sel_mdl
#save(sav_fin_mdl, sav_sel_mdl, file=paste0(glb_out_pfx, "sav_mdl.RData"))
# load(file=paste0(glb_out_pfx, "sav_mdl_01.RData"), verbose=TRUE)
# prv_fin_mdl <- sav_fin_mdl; prv_sel_mdl <- sav_sel_mdl
# load(file=paste0(glb_out_pfx, "sav_mdl.RData"), verbose=TRUE)
# cur_fin_mdl <- sav_fin_mdl; cur_sel_mdl <- sav_sel_mdl
# all.equal(cur_fin_mdl, prv_fin_mdl)
# cur_fitobs_df <- cur_fin_mdl$trainingData; prv_fitobs_df <- prv_fin_mdl$trainingData; all.equal(cur_fitobs_df, prv_fitobs_df)
# nrow(cur_fitobs_df); nrow(prv_fitobs_df)
# names(cur_fitobs_df); names(prv_fitobs_df)
# all.equal(cur_fin_mdl$bestTune, prv_fin_mdl$bestTune)
# all.equal(glb_sel_mdl, sav_sel_mdl)
# cur_fitobs_df <- cur_sel_mdl$trainingData; prv_fitobs_df <- prv_sel_mdl$trainingData; all.equal(cur_fitobs_df, prv_fitobs_df)
# head(myget_feats_importance(glb_sel_mdl)); head(myget_feats_importance(sav_sel_mdl))
# head(myget_feats_importance(cur_sel_mdl)); head(myget_feats_importance(prv_sel_mdl))
# tmp_replay_lst <- replay.petrisim(pn=glb_analytics_pn,
# replay.trans=(glb_analytics_avl_objs <- c(glb_analytics_avl_objs,
# "data.new.prediction")), flip_coord=TRUE)
# print(ggplot.petrinet(tmp_replay_lst[["pn"]]) + coord_flip())
glb_chunks_df <- myadd_chunk(glb_chunks_df, "display.session.info", major.inc=TRUE)## label step_major step_minor label_minor bgn end
## 16 predict.data.new 8 0 0 541.250 561.302
## 17 display.session.info 9 0 0 561.302 NA
## elapsed
## 16 20.052
## 17 NANull Hypothesis (\(\sf{H_{0}}\)): mpg is not impacted by am_fctr.
The variance by am_fctr appears to be independent. #{r q1, cache=FALSE} # print(t.test(subset(cars_df, am_fctr == "automatic")$mpg, # subset(cars_df, am_fctr == "manual")$mpg, # var.equal=FALSE)$conf) # We reject the null hypothesis i.e. we have evidence to conclude that am_fctr impacts mpg (95% confidence). Manual transmission is better for miles per gallon versus automatic transmission.
## label step_major step_minor label_minor bgn
## 5 extract.features 3 0 0 47.050
## 11 fit.models 6 1 1 409.619
## 10 fit.models 6 0 0 342.849
## 7 cluster.data 3 2 2 280.448
## 14 fit.data.training 7 0 0 504.742
## 12 fit.models 6 2 2 477.215
## 16 predict.data.new 8 0 0 541.250
## 1 import.data 1 0 0 23.954
## 9 select.features 5 0 0 331.640
## 15 fit.data.training 7 1 1 533.211
## 13 fit.models 6 3 3 498.303
## 2 inspect.data 2 0 0 40.783
## 8 partition.data.training 4 0 0 330.238
## 3 scrub.data 2 1 1 45.357
## 4 transform.data 2 2 2 46.439
## 6 manage.missing.data 3 1 1 280.282
## end elapsed duration
## 5 280.281 233.231 233.231
## 11 477.214 67.595 67.595
## 10 409.618 66.769 66.769
## 7 330.237 49.789 49.789
## 14 533.210 28.468 28.468
## 12 498.302 21.087 21.087
## 16 561.302 20.052 20.052
## 1 40.783 16.829 16.829
## 9 342.849 11.209 11.209
## 15 541.249 8.038 8.038
## 13 504.741 6.438 6.438
## 2 45.356 4.574 4.573
## 8 331.639 1.401 1.401
## 3 46.438 1.081 1.081
## 4 47.049 0.610 0.610
## 6 280.448 0.166 0.166
## [1] "Total Elapsed Time: 561.302 secs"
## label step_major
## 9 extract.features_bind.DXM 8
## 5 extract.features_build.corpus 4
## 7 extract.features_report.DTM 6
## 8 extract.features_bind.DTM 7
## 3 extract.features_process.text 3
## 6 extract.features_extract.DTM 5
## 2 extract.features_factorize.str.vars 2
## 1 extract.features_bgn 1
## 4 extract.features_process.text_reporting_compound_terms 3
## step_minor label_minor bgn end elapsed duration
## 9 0 0 80.382 274.829 194.447 194.447
## 5 0 0 48.741 71.464 22.723 22.723
## 7 0 0 72.831 77.561 4.731 4.730
## 8 0 0 77.562 80.382 2.820 2.820
## 3 0 0 47.145 48.734 1.589 1.589
## 6 0 0 71.465 72.831 1.366 1.366
## 2 0 0 47.074 47.144 0.071 0.070
## 1 0 0 47.057 47.074 0.017 0.017
## 4 1 1 48.735 48.740 0.005 0.005
## [1] "Total Elapsed Time: 274.829 secs"